Thoracic radiation therapy is a cornerstone in the treatment of malignancies such as breast cancer, lung cancer, esophageal cancer, and lymphoma. While its long-term cardiac risks are well known, there is limited data on how prior thoracic irradiation is associated with outcomes in patients hospitalized with acute myocardial infarction (AMI). This study evaluates the association between prior chest irradiation and in-hospital outcomes among patients admitted with AMI. A retrospective cohort study using the National Inpatient Sample (2016–2022) was conducted. Adult AMI admissions were identified via ICD-10 codes and stratified by history of thoracic radiation. Multivariable regression and propensity matching were used to evaluate the association of prior radiation on in-hospital mortality (primary outcome), and secondary outcomes including hospitalization cost, length of stay, and use of intensive interventions. Of 4,353,204 AMI hospitalizations, 5,280 had a history of thoracic radiation. Prior radiation was associated with increased in-hospital mortality (aOR: 1.55, 95% CI 1.06 to 2.27, p = 0.023). There were no significant differences in hospitalization cost (−$6,126, p = 0.196) or length of stay (−0.20 days, p = 0.327). Patients with prior radiation were more likely to have do-not-resuscitate orders (aOR: 2.15, p <0.001) and receive palliative care consultations (aOR: 2.43, p <0.001). Prior thoracic radiation is associated with worse in-hospital survival following AMI, along with greater palliative involvement and end-of-life care decisions. These findings underscore the need for cardio-oncology–informed inpatient care in this high-risk population.
Radiation therapy to the thoracic region remains a fundamental component in the treatment of various malignancies, including breast cancer and thoracic malignancies, such as nonsmall cell lung cancer (NSCLC) and lymphoma. For patients with left-sided breast cancer, radiation is often delivered to the chest wall and regional lymph nodes, including the internal mammary, while in Hodgkin lymphoma and other mediastinal tumors, treatment volumes frequently encompass the lungs and cardiac structures. , As these treatments often involve critical thoracic anatomy, many patients with prior malignancy remain at risk for long-term complications. In this study, we use the term “cancer survivor” to refer to individuals with a history of thoracic malignancy who are alive at the time of hospitalization, regardless of current treatment status. Although such treatments have dramatically improved cancer-specific survival, they are not without long-term cardiovascular consequences.
Thoracic radiation has been associated with an elevated risk of cardiovascular complications, particularly coronary artery disease. One of the most well-documented adverse effects of thoracic radiation is the development of radiation-induced heart disease (RIHD), which encompasses a spectrum of conditions such as pericarditis, cardiomyopathy, valvular disease, and most notably, coronary artery disease (CAD). Potential contributors to this increased risk include radiation-induced microvascular dysfunction, endothelial cell damage, myocardial fibrosis, and accelerated atherosclerosis. Over time, these effects impair cardiac function, often in patients who may otherwise have long-term survival after cancer therapy. In a comprehensive review, Lee et al summarizes the spectrum of radiation induced heart disease, highlight endothelial and microvascular injury as central mechanisms, and recommend longitudinal cardiology surveillance after chest radiotherapy.
Although there have been improvements in radiation delivery, such as intensity-modulated radiotherapy (IMRT), deep inspiration breath hold (DIBH), and intensity-modulated proton therapy (IMPT), depending on the location of the tumor, meeting the dose constraints to the cardiac structures may not be feasible, as the trade-off between achieving planning target volume (PTV) coverage and protecting organs at risk is of great importance when developing a treatment plan. This could be seen in left-sided breast cancer and mediastinal lymphoma but not exclusively. , The dose-response relationship between cardiac radiation exposure and coronary events is well established. A pivotal study demonstrated that for every 1 Gy increase in mean heart dose, the risk of major coronary events, including myocardial infarction (MI), coronary revascularization, or cardiac death, increases by 7.4%, with no apparent safe threshold, also known as deterministic effect. Importantly, these risks are not confined to high-dose treatment, as a mean dose >15 Gy to the left coronary artery increases the risk of cardiovascular events, , as endothelial damage is thought to be the primary culprit of cardiac toxicity, in contrast when in the past was associated to pericarditis, when newer techniques such as IMRT/IMPT were not widely available. These risks manifest several years after radiation, especially in patients with excellent prognosis and overall survival, such as in patients treated for lymphoma that received systemic therapy (doxorubicin). Notably, the risk of ischemic heart disease increases within 5 years of treatment and continues for at least 2 decades thereafter.
Despite the known association between prior thoracic radiation and cardiovascular disease, there is a lack of data on how radiation to the mediastinum is related to the outcomes of patients hospitalized with acute myocardial infarction (AMI). It remains unclear whether patients who were treated with radiation therapy experience worse in-hospital outcomes, incur higher healthcare costs, or face prolonged hospitalization during AMI admissions.
To address this gap, we used a large, nationally representative inpatient database to evaluate the association of prior mediastinal irradiation on clinical outcomes among patients hospitalized with AMI. We specifically examined in-hospital mortality, total cost of hospitalization, and length of stay. Our aim was to provide insights into how prior radiation exposure might correlate with the acute course and resource utilization during AMI admissions and to inform risk stratification and management strategies for this growing high-risk patient population.
Methods
Using data from the National Inpatient Sample (NIS) between 2016 and 2022, a retrospective cohort study was conducted.
The NIS is a nationally representative database, as part of the greater family of databases created for the Healthcare Cost and Utilization Project (HCUP). Released annually by the Agency for Healthcare Research and Quality (AHRQ), this database approximates a 20-percent stratified sample of discharges from hospitals across the United States. A range of information is provided about patients, including demographic characteristics, procedures performed, total hospitalization costs, in-hospital outcomes, and primary and secondary diagnoses. Diagnoses are stratified with a primary ICD-10 code followed by multiple secondary ICD-10 codes. Secondary diagnoses cannot be discerned from those that began during the hospitalization and those that began prior. Furthermore, the NIS provides discharge weights, enabling researchers to extrapolate sample discharges to produce nationally representative estimates. Because the NIS dataset is de-identified, both informed consent and Institutional Review Board approval were not required.
All patients from the National Inpatient Database from the year 2016 to 2022 were analyzed. Baseline characteristics were initially stratified, including race, age, gender, and hospital characteristics. Patients who were 18 or older at the time of admission were then further stratified. Using ICD-10 codes beginning with I21, all patients admitted with AMI were identified. To identify a history of prior thoracic irradiation, ICD code Z92.3 was used to identify individuals with a history of prior irradiation and further refined by identifying ICD-10 codes for malignancies of the thorax and chest wall (Supplementary Table 1). For the purposes of this study, we defined cancer “survivors” as patients with a prior history of thoracic malignancy who were alive at the time of hospitalization, regardless of ongoing cancer treatment. This reflects a pragmatic definition consistent with population-level survivorship research and aligns with the structure of the NIS database, which lacks treatment timeline detail. Patients admitted primarily for MI were stratified into 2 groups based on the presence of a history of prior thoracic irradiation. All outcomes of choices were identified based on the respective ICD-10 codes for each outcome ( Figure 1 ).
Flowchart of study population selection.
Diagram illustrating the inclusion and exclusion criteria used to define the final study cohort from the initial patient population.
All statistical analyses were performed using Stata version 18. To minimize confounding, 1:1 nearest-neighbor propensity score matching without replacement was conducted using a caliper width of 0.2. Propensity scores were estimated using multivariable logistic regression based on demographic, hospital, and clinical variables, including age, sex, race, ZIP code–based income quartile, Charlson comorbidity index category, hospital teaching status, region, bed size, insurance type, and comorbidities such as congestive heart failure, arrhythmia, valvular disease, pulmonary hypertension, peripheral vascular disease, hypertension, chronic obstructive pulmonary disease, diabetes, liver disease, hyperlipidemia, alcohol abuse, and prior chemotherapy exposure ( Table 1 ).
Table 1
Baseline demographic and hospital characteristics prior to propensity match
| Demographic | AMI without thoracic irradiation ( n = 4,347,924) | AMI with thoracic irradiation ( n = 5,280) | p-value |
|---|---|---|---|
| Sex (%) | <0.001 | ||
| Female | 37.2% | 52.9% | |
| Male | 63.0% | 47.1% | |
| Mean age | 66.8 | 71.4 | <0.001 |
| Race | <0.001 | ||
| White | 73.0% | 78.6% | |
| Black | 11.4% | 11.8% | |
| Hispanic | 9.0% | 4.6% | |
| Asian or Pacific Islander | 3.0% | 2.4% | |
| Native American | 0.6% | 0.5% | |
| Other | 3.0% | 2.2% | |
| National quartile for median household income | 0.027 | ||
| 0–25th percentile | 30.7% | 27.0% | |
| 26th–50th percentile | 27.5% | 27.1% | |
| 51st–75th percentile | 23.4% | 25.3% | |
| 76th–100th percentile | 18.4% | 20.6% | |
| Payer | <0.001 | ||
| Medicare | 56.3% | 73.8% | |
| Medicaid | 9.8% | 7.1% | |
| Private insurance | 25.8% | 15.8% | |
| Self-pay | 4.7% | 1.0% | |
| No charge | 0.4% | 0.1% | |
| Other | 3.1% | 2.2% | |
| Hospital region | <0.001 | ||
| Northeast | 17.1% | 15.0% | |
| Midwest | 22.3% | 29.7% | |
| South | 41.4% | 34.1% | |
| West | 19.2% | 21.2% | |
| Hospital bed size | 0.443 | ||
| Small | 18.9% | 17.9% | |
| Medium | 30.3% | 31.9% | |
| Large | 50.8% | 50.2% | |
| Hospital location/teaching status | <0.001 | ||
| Rural | 7.7% | 6.8% | |
| Urban/Nonteaching | 30.0% | 16.4% | |
| Urban/Teaching | 71.4% | 76.8% | |
| History of congestive heart failure | 43.1% | 45.4% | 0.130 |
| History of arrhythmia | 36.6% | 38.1% | 0.325 |
| History of valvular disease | 14.7% | 16.0% | 0.217 |
| History of pulmonary hypertension | 6.0% | 7.9% | 0.007 |
| History of peripheral vascular disease | 11.7% | 18.0% | <0.001 |
| History of hypertension | 81.8% | 81.7% | 0.953 |
| History of chronic obstructive pulmonary disease | 21.4% | 42.9% | <0.001 |
| History of diabetes | 40.6% | 33.1% | <0.001 |
| History of liver disease | 4.3% | 3.3% | 0.103 |
| History of hyperlipidemia | 68.6% | 62.7% | <0.001 |
| History of alcohol abuse | 0.4% | 0.4% | 0.943 |
| Prior chemotherapy exposure | 0.8% | 46.7% | <0.001 |
Comparison of baseline variables including age, sex, race/ethnicity, median household income quartile, insurance type, hospital region, hospital size, teaching status, and comorbidities between patients with and without prior thoracic radiation.
After matching, covariate balance was assessed using standardized mean differences and variance ratios, with all variables demonstrating adequate balance (mean bias 2.7%, all standardized differences <10%) (Supplementary Table 1). Doubly robust multivariable logistic and linear regression models were then applied to the matched cohort, adjusting for all baseline covariates used in the propensity score model. Logistic regression was used for binary outcomes, and linear regression was used for continuous outcomes.
In a secondary analysis, outcomes were compared between patients with prior radiation exposure alone and those with chemotherapy exposure alone. A separate propensity score model was constructed using the same covariates, followed by 1:1 matching without replacement. Binary and continuous outcomes were evaluated using both unadjusted and adjusted regression models. All odds ratios reported reflect models adjusted for the full post–covariate match covariate set.
To evaluate whether the association between treatment exposure and outcomes varied by malignancy type, interaction analyses were performed in both the primary (radiation vs no radiation) and secondary (radiation vs chemotherapy) comparisons. Patients were categorized into the most common subgroups: lung/airway, breast, esophageal, and laryngeal cancers. All remaining intrathoracic malignancies (including thyroid, thymic, mediastinal, and lymphoid tumors, as well as other less frequent categories) were pooled into a single group to preserve model stability. Cancer type was incorporated into multivariable models as an interaction term with exposure, and stratified estimates were generated for the major cancer groups. When a statistically significant interaction was identified, follow-up models were used to determine which cancer subtypes accounted for the observed effect.
The primary outcome was in-hospital mortality. Secondary outcomes included total hospital charges, length of stay (LOS), ischemic stroke, acute kidney injury, cardiac arrest, cardiopulmonary resuscitation (CPR), blood transfusion, use of vasopressors, mechanical circulatory support, mechanical ventilation, palliative consultations, and Do Not Resuscitate (DNR) orders. Patients with missing variables were excluded from the analysis. p-values less than 0.05 were considered statistically significant.
Results
A total of 4,353,204 patients were identified who were admitted for a primary diagnosis of AMI. Of these patients, 5,280 were identified as having previous thoracic irradiation. Compared to patients without prior thoracic irradiation, those with a history of irradiation were significantly older (mean age 71.4 vs 66.8 years, p <0.001) and more likely to be female (52.9% vs 37.2%, p <0.001) ( Table 1 ). Racial distribution differed across groups, with a higher proportion of White patients in the irradiated cohort (78.6% vs 73.0%) and a lower proportion identifying as Hispanic (4.6% vs 9.0%) or Other (2.2% vs 3.0%) (p <0.001). Socioeconomic status also varied modestly: patients with prior irradiation were less likely to fall within the lowest income quartile (27.0% vs 30.7%) and more frequently in the top quartile (20.6% vs 18.4%, p = 0.027). Insurance status differed significantly between groups, with Medicare coverage more common in the irradiated cohort (73.8% vs 56.3%, p <0.001), while privately insured and self-pay patients were more prevalent in the nonirradiated group.
In terms of regional distribution, patients with a history of thoracic irradiation were more likely to be hospitalized in the Midwest (29.7% vs 22.3%) and less likely in the South (34.1% vs 41.4%, p <0.001). While the distribution of hospital bed sizes did not differ between groups (p = 0.443), patients with prior irradiation were more often treated at urban teaching hospitals (76.8% vs 71.4%) and less frequently at urban nonteaching facilities (16.4% vs 30.0%, p <0.001). Comorbidity profiles showed similar rates of congestive heart failure, arrhythmias, valvular disease, hypertension, and liver disease. However, irradiated patients had higher rates of pulmonary hypertension (7.9% vs 6.0%, p = 0.007) and peripheral vascular disease (18.0% vs 11.7%, p <0.001), but lower rates of COPD (42.9% vs 21.4%, p <0.001), diabetes (33.1% vs 40.6%, p <0.001), and hyperlipidemia (62.7% vs 68.6%, p <0.001). Prior chemotherapy exposure was notably more common in the irradiated cohort (46.7% vs 0.8%, p <0.001). Among patients with a history of thoracic irradiation, lung and airway malignancies were the most common (67.3%), followed by breast (20.5%) and esophageal cancers (5.7%) ( Table 4 ).
Table 4
Frequency of thoracic malignancy types in patients with a history of thoracic malignancy admitted with acute myocardial infarction
| C ancer type | Percent (%) |
|---|---|
| Lung/Airway | 67.3% |
| Breast | 20.5% |
| Esophagus | 5.7% |
| Larynx | 2.8% |
| Thyroid | 0.7% |
| Heart/Mediastinum | 0.4% |
| Thymus | 0.1% |
| Lymphoma | 0.3% |
| Other | 2.3% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
