Is the Lifetime Malignancy Risk in United States Military Personnel Sustaining Combat-related Trauma Increased Because of Radiation Exposure From Diagnostic Imaging?

Abstract

Background

Patients with complex polytrauma in the military and civilian settings are often exposed to substantial diagnostic medical radiation because of serial imaging studies for injury diagnosis and subsequent management. This cumulative radiation exposure may increase the risk of subsequent malignancy. This is particularly true for combat-injured servicemembers who receive care at a variety of facilities worldwide. Currently, there is no coordinated effort to track the amount of radiation exposure each servicemember receives, nor a surveillance program to follow such patients in the long term. It is important to assess whether military servicemembers are exposed to excessive diagnostic radiation to mitigate or prevent such occurrences and monitor for carcinogenesis, when necessary. The cumulative amount of radiation exposure for combat-wounded and noncombat-wounded servicemembers has not been described, and it remains unknown whether diagnostic radiation exposure meets thresholds for an increased risk of carcinogenesis.

Questions/purposes

We performed this study to (1) quantify the amount of exposure for combat-wounded servicemembers based on medical imaging in the first year after injury and compare those exposures with noncombat-related trauma, and (2) determine whether the cumulative dose of radiation correlates to the Injury Severity Score (ISS) across the combat-wounded and noncombat-wounded population combined.

Methods

We performed a retrospective study of servicemembers who sustained combat or noncombat trauma and were treated at Walter Reed National Military Medical Center from 2005 to 2018. We evaluated patients using the Department of Defense Trauma Registry. After consolidating redundant records, the dataset included 3812 unique servicemember encounters. Three percent (104 of 3812) were excluded because of missing radiation exposure data in the electronic medical record. The final cohort included 3708 servicemembers who had combat or noncombat injury trauma, with a mean age at the time of injury of 26 ± 6 years and a mean ISS of 18 ± 12. The most common combat trauma mechanisms of injury were blast (in 65% [2415 of 3708 patients]), followed by high-velocity gunshot wounds (in 22% [815 of 3708 patients]). We calculated the cumulative diagnostic radiation dose exposure at 1 year post-traumatic injury in patients with combat-related trauma and those with noncombat trauma. We did this by multiplying the number of imaging studies by the standardized effective radiation dose for each imaging study type. We then performed analysis of variance for four data subsets (battle combat trauma, nonbattle civilian trauma, high ISS, and high radiation exposure [> 50 mSv]) independently. To evaluate whether the total number of imaging studies, radiation exposure, and ISS values differed between battle-wounded and nonbattle-wounded patients, we performed a pairwise t-test.

Results

The mean radiation exposure for combat-related injuries was 35 ± 26 mSv while the mean radiation exposure for noncombat-related injuries was 22 ± 33 mSv in the first year after injury. In the first year after trauma, 44% of patients (1626 of 3708) were exposed to high levels of radiation that were greater than 20 mSv, and 23% (840 of 3708) were exposed to very high levels of radiation that were greater than 50 mSv. Servicemembers with combat trauma-related injuries had eight more imaging studies than those who sustained noncombat injuries. Servicemembers with combat trauma injuries (35 ± 26 mSv) were exposed to more radiation (approximately 4 mSv) than patients treated for noncombat injuries (22 ± 33 mSv) (p = 0.01). We found that servicemembers with combat injuries had a higher ISS than servicemembers with noncombat trauma (p < 0.001). We found a positive correlation between radiation exposure and ISS for servicemembers. The positive relationship between radiation exposure and ISS held for combat trauma (r² = 0.24; p < 0.001), noncombat trauma (r² = 0.20; p < 0.001), servicemembers with a high ISS (r² = 0.10; p < 0.001), and servicemembers exposed to high doses of radiation (r² = 0.09; p < 0.001).

Conclusion

These data should be used during clinical decision-making and patient counseling at military treatment facilities and might provide guidance to the Defense Health Agency. These recommendations will help determine whether the benefits of further imaging outweigh the risk of carcinogenesis. If not, we need to develop interdisciplinary clinical practice guidelines to reduce or minimize radiation exposure. It is important for treating physicians to seriously weigh the risk and benefits of every imaging study ordered because each test does not come without a cumulative risk.

Level of Evidence

Level III, therapeutic study.

Introduction

The mean radiation exposure in an individual not sustaining trauma is 2.4 mSv/year [4, 10]. Medical radiation exposure after trauma is increasing despite standardized care patterns and treatment algorithms. Previous studies suggested there is an increased cancer risk with exposures of 20 to 40 mSv per year [4, 10]. Evidence suggests acute radiation exposure greater than 5 mSv increases the risk of carcinogenesis. This risk is substantially increased when dose exposure is greater than 50 mSv [2, 5]. These risks exist for acute and chronic exposures, particularly in doses of greater than 100 mSv [1, 5]. The International Commission on Radiological Protection estimated a 4% to 5% increased relative risk of fatal cancer after an average person receives a cumulative whole-body radiation dose of 1 Sv (1 Sv = 1000 mSv) [7].

In the United States military, patients with combat trauma are treated at a variety of facilities while en route to definitive treatment. Thus, these patients are often exposed to substantial medical radiation because of serial imaging studies for injury diagnosis and management. Patients with civilian trauma, especially those with orthopaedic conditions, are likewise exposed to high amounts of radiation [8]. However, there is no coordinated effort to track the amount of radiation exposure each patient receives, nor any current long-term surveillance program to follow patients. It is unknown whether an individual patient’s cumulative diagnostic radiation exposure increases the risk of carcinogenesis.

We performed this study to (1) quantify the amount of exposure for combat-wounded servicemembers based on medical imaging in the first year after injury and compare those exposures with noncombat-related trauma, and (2) determine whether the cumulative dose of radiation correlates to the Injury Severity Score (ISS) across the combat-wounded and noncombat-wounded population combined.

Patients and Methods

Data Source

This retrospective, observational study followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for reporting an accurate and complete observational study [11]. We queried the Department of Defense Trauma Registry, which tracks the epidemiology, treatments, and outcomes of patients receiving care for combat casualties. The Department of Defense Trauma Registry was used because it longitudinally collects United States military beneficiary data on demographics, mechanism of injury, diagnoses, treatments, and outcome of combat and noncombat injuries. The Department of Defense Trauma Registry is the most comprehensive military database from the point of injury to the final disposition. Registry abstractors are required to maintain a 90% accuracy rate as a minimum standard [9].

Demographic Data and Sample Size

The Department of Defense Trauma Registry was queried for all trauma servicemembers transported from definitive care to Walter Reed National Military Medical Center from 2005 to 2018. After removal of duplications, this yielded a dataset of 3812 unique servicemembers. Of these patients, 3% (104 of 3812) were missing medical rational data and were excluded from the analysis. Our final dataset included 3708 servicemembers with a mean age of 26 ± 6 years and a mean ISS of 18 ± 12 (Fig. 1). Most injuries were combat-related, occurring in 84% (3102 of 3708) of the population. Sixty-five percent of the patients (2415 of 3708) had blast injuries, and 22% (815 of 3708) experienced high-velocity gunshot wounds (Table 1).

Fig. 1.

This figure shows the distribution of the ISS, where red lines indicate separation between injury categories. Also shown, in blue, are the mean, standard errors, and regression model of the radiation exposure (in mSv) per injury severity scores. A color image accompanies the online version of this article.

Table 1.

Demographics of the study population

Variable	Type of trauma		Type of injury				ISS score > 15 (n = 1931)	Radiation exposure		Combat versus noncombat trauma, mean difference (95% CI) and p value
	Combat (n =3102)	Noncombat (n = 606)	Penetrating (n = 2550)	Blunt (n = 1145)	Burn (n = 11)	Other (n = 2)		mSv > 20 (n = 1626)	mSv > 50 (n = 840)
Age in years	25 ± 6	27 ± 8	26 ± 6	26 ± 7	23 ± 3	30 ± 2	26 ± 6	26 ± 6	26 ± 6
ISS score	19 ± 12	15 ± 11	19 ± 12	17 ± 12	11 ± 10	9 ± 0	27 ± 10	25 ± 12	29 ± 12	3.63 (2.64 to 4.63) p < 0.001
Number of imaging studies	29 ± 28	21 ± 22	29 ± 27	26 ± 26	16 ± 22	24 ± 28	39 ± 31	47 ± 29	62 ± 31	8.37 (6.35 to 10.39) p < 0.001
Diagnostic radiation exposure in mSv	35 ± 26	22 ± 33	34 ± 45	30 ± 43	19 ± 27	38 ± 22	52 ± 52	67 ± 49	99 ± 50	3.60 (0.63 to 6.57) p = 0.01

Data are presented as mean ± standard deviation unless indicated otherwise; ISS = Injury Severity Score.

We then calculated the sample size needed for subsequent analyses using the “pwr” v1.3-0 package in R. We designated the significance level at 0.05, the power at 0.90, and the effect size at 0.20. This resulted in a sample size of 526 servicemembers per group.

Study Design and Variables

We derived nine variables from the total Department of Defense Trauma Registry dataset that included ISS, age, race, military branch, whether the injury was combat-related or noncombat-related, injury mechanism, injury type, total number of imaging studies and type undergone during the first year after injury, and the total medical radiation exposure (in mSv) (Supplemental Table 1; http://links.lww.com/CORR/A979). The variables of ISS, total radiation exposure, and total number of imaging studies did not follow a normal distribution. Therefore, we normalized the data using the min-max normalization method, which allowed us to perform parametric statistical analyses on these data.

Statistical Analysis

Radiation Exposure

From the patients’ electronic medical records, we calculated the total medical radiation exposure by multiplying the number of imaging studies by the mean radiation dose per scan [6]. This enabled us to identify patients who had high radiation exposure, defined as levels greater than 20 mSv of medical radiation, and those with very high radiation exposure, defined as levels greater than 50 mSv [3].

Comparing Exposure of Military and Noncombat Trauma

To determine whether the total number of imaging studies, radiation exposure, and ISS values differed between servicemembers with combat-related trauma and those with noncombat-related trauma, we performed a pairwise t-test. All reported p values are two sided with the significance level set at < 0.05.

Correlation Between Radiation Exposure and ISS

To understand how the number of imaging studies undergone by a patient and radiation exposure (in mSv) are related to the ISS, we built a linear model using the ISS value as the independent variable and the total number of imaging studies and radiation exposure as the dependent variables, separately for each of the subgroups (combat, noncombat, high ISS, and high radiation exposure). A similar approach was used to investigate how the number of imaging studies undergone by a patient and radiation exposure are related to patient age.

Additionally, analyses of variance were used with Turkey post hoc tests to determine whether there were differences in the categories that populated the mechanisms of injury, injury type, military branch, or race variables, as well as in the total number of imaging studies and radiation exposure. We performed these analyses on the four data subsets (combat, noncombat, high ISS, and high radiation exposure) independently. All analyses were performed using R statistical programming language.

Results

The mean radiation exposure for combat-related injuries was 35 ± 26 mSv, while the mean radiation exposure for noncombat-related injuries was 22 ± 33 mSv in the first year after injury. In the first year after injury, 44% (1626 of 3708 patients) had high radiation exposure (defined as > 20 mSv) and 23% (840 of 3708 patients) had very high radiation exposure (defined as > 50 mSv).

Patients with combat-related injuries had a higher ISS (19 ± 12) than those with noncombat injuries (15 ± 11; p < 0.001) (Table 1). This difference in ISS also led to a difference in the total number of imaging studies; patients with combat-related injuries (29 ± 28) had, on average, eight more imaging studies than patients with noncombat injuries did (21 ± 22) (p < 0.001) (Table 1). Subsequently, patients with battle injuries were exposed to more radiation by an average of 4.0 mSv (p = 0.01) (Table 1).

We found a positive correlation between radiation exposure and ISS for servicemembers. The positive relationship between radiation exposure and ISS held for combat trauma (r² = 0.24; p < 0.001) (Fig. 2A), noncombat trauma (r² = 0.20; p < 0.001) (Fig. 2B), servicemembers with a high ISS (r² = 0.10; p < 0.001) (Fig. 2C), and servicemembers exposed to high doses of radiation (r² = 0.09; p < 0.001) (Fig. 2D). We found no difference in the amount of radiation exposure based on mechanism of injury.

Fig. 2.

These regression models represent the correlation between ISS (x-axis) and radiation exposure (y-axis) for servicemembers who experienced (A) combat trauma, (B) noncombat trauma, (C) high ISS values, and (D) exposure to high radiation. A color image accompanies the online version of this article.

Discussion

Servicemembers with combat and noncombat-related injuries undergo serial diagnostic radiation exposure en route to the military treatment facility for definitive management. The cumulative risks associated with serial diagnostic imaging studies are important to understand. Estimating the risk of carcinogenesis based on imaging examinations is controversial given the potential harm from ionizing radiation and limited epidemiologic data. However, there is increasing consensus in the medical and scientific communities that the risk, however small, is real and may increase the risk of subsequent fatal malignancy. We found that service members with combat-related injuries were exposed to higher radiation doses, had higher ISS, and had more imaging studies than servicemembers with noncombat-related injuries. Additionally, in the first year, almost half of servicemembers were exposed to high doses of radiation, and almost one-fourth of service members were exposed to very high doses of radiation. Lastly, we found a positive correlation between radiation exposure and ISS for servicemembers with combat-related trauma or noncombat-related trauma and servicemembers with a high ISS.

Limitations

First, the Department of Defense Trauma Registry inherently depends on the quality of the data input. Registry abstractors are required to maintain a 90% accuracy rate as the minimum standard [9]. To limit errors, the registry data undergo rigorous periodic audits and continuous quality review [9]. Additionally, the registry is an incidence registry, meaning that a single patient may have multiple injury encounters and therefore registry records. This may lead to data discrepancies or redundancy. For our analysis, duplicated registry records were consolidated without apparent record redundancy. This did not impact our final study cohort size or the power of our analysis. Next, the registry focuses on combat trauma; however, 17% (seven of 42) of the contributing military treatment facilities also submit data on noncombat trauma. Four of these seven facilities are Army military treatment facilities. This did not impact our final analysis because the military treatment facilities follow the same patterns of care regardless of service branch. Next, we were unable to obtain patient-specific data such as height or weight at the time of injury and/or diagnostic radiation exposures. Thus, the mean radiation doses were based on American College of Radiology effective doses in radiology and diagnostic nuclear medicine [6]. Our methodology thus simplifies a highly complex topic; the American College of Radiology effective doses are values calculated for an average-sized adult [6]. This allowed us to describe the generalized risk of carcinogenesis for these patients. We acknowledge this may limit the accuracy of individualized or organ-specific risk.

The mean radiation exposure for combat-related injuries was 35 ± 26 mSv, while the mean radiation exposure for noncombat-related injuries was 22 ± 33 mSv in the first year after injury. Servicemembers with combat-related trauma injuries were exposed to more radiation (approximately 4 mSv) than servicemembers treated for noncombat injuries. The radiation exposure in combat-injured and noncombat-injured servicemembers was comparable to the mean radiation exposure of 30 mSv reported by Howard et al. [5] in civilian patients with polytrauma, with a mean ISS of 29. The mean ISS for servicemembers with combat-related and noncombat-related trauma were 19 and 15, respectively. Conversely, some patients with high ISSs underwent fewer imaging studies and therefore were exposed to less radiation. It is unclear why patients with a high ISS did not receive more imaging studies than patients with less-severe injuries. However, radiation exposure was not associated with patient demographics; patients were treated equally regardless of age, race, military service branch, and injury type.

In our patient cohort, 44% (1626 patients) had high radiation exposure and 23% (840 patients) had very high radiation exposure, and radiation exposure directly correlated with ISS. There is evidence to suggest the risk of carcinogenesis substantially increases when the dose exposure is greater than 50 mSv [2, 5]. Annually, a person receives a mean radiation dose of approximately 2.4 mSv from background radiation [4, 10]. Our findings suggest some patients are receiving nearly 10 to 20 times the mean annual dose. Consistent with reported estimates from civilian trauma centers [5], this study found that with more severe injury, there was greater radiation exposure, resulting in an anticipated increased risk of carcinogenesis [2, 8].

We must be cognizant of the implications of every clinical decision we make when evaluating patients. As physicians, we need to recognize that imaging studies, particularly serial examinations and large-field CT scans, do not come without an inherent risk. Thus, it is imperative we weigh the risk and benefits before ordering another ionizing radiation examination. Furthermore, it is important to counsel patients sustaining either combat or noncombat trauma-related injuries that they may have an increased risk of subsequent carcinogenesis. To appropriately counsel patients, interdisciplinary clinical practices should be developed to provide guidance on surveillance for patients with high or very high levels of exposure.

Conclusion

In a United States military healthcare population, we found that nearly half of the patients were subjected to high levels of diagnostic radiation exposure and nearly one-fourth received very high levels, potentially increasing the relative risk of fatal cancer. Clinicians must be thoughtful and remember the risk associated with radiation exposure from imaging studies. In military and civilian trauma systems, greater emphasis is required on forward transfer of performed imaging and avoiding imaging studies at the transient, transferring facility that are not required for immediate management in order to prevent unnecessary study duplication. During clinical decision-making, we recommend that each patient be evaluated individually when determining whether the benefits of further imaging outweigh the risk of carcinogenesis. Interdisciplinary clinical practice guidelines might address practical measures to make diagnostic radiation exposure as low as reasonably achievable. We further advocate for institutional radiology registries in the electronic medical record that would monitor patients for acute and chronic cumulative radiation exposure levels to prospectively identify patients at risk and notify clinicians ordering another scan. Future studies should seek to address such interventions or include long-term follow-up of patients with trauma (or those who have experienced cancer, who may also be exposed to high effective doses of diagnostic radiation) to improve our epidemiologic understanding of malignancy risk.