Cardiovascular risk factors, radiation therapy, and myocardial infarction among lymphoma survivors

Abstract

Background

Mediastinal radiation is associated with increased risk of myocardial infarction (MI) among non-Hodgkin lymphoma (NHL) survivors.

Objective

To evaluate how preexisting cardiovascular risk factors (CVRFs) modify the association of mediastinal radiation and MI among a national population of NHL survivors with a range of CVRFs.

Material and methods

Using Danish registries, we identified adults diagnosed with lymphoma 2000–2010. We assessed MI from one year after diagnosis through 2016. We ascertained CVRFs (hypertension, dyslipidemia, and diabetes), vascular disease, and intrinsic heart disease prevalent at lymphoma diagnosis. We used multivariable Cox regression to test the interaction between preexisting CVRFs and receipt of mediastinal radiation on subsequent MI.

Results

Among 3151 NHL survivors (median age 63, median follow-up 6.5 years), 96 were diagnosed with MI. Before lymphoma, 32% of survivors had ≥1 CVRF. 8.5% of survivors received mediastinal radiation. In multivariable analysis, we found that mediastinal radiation (HR = 1.96; 95% CI = 1.09–3.52), and presence of ≥1 CVRF (HR = 2.71; 95% CI = 1.77–4.15) were associated with an increased risk of MI. Although there was no interaction on the relative scale (p = 0.14), we saw a clinically relevant absolute increase in risk for patients with CVRF from 10-year of MI of 10.5% without radiation to 29.5% for those undergoing radiation.

Conclusion

Patients with CVRFs have an importantly higher risk of subsequent MI if they have mediastinal radiation. Routine evaluation of CVRFs and optimal treatment of preexisting cardiovascular disease should continue after receiving cancer therapy. In patients with CVRFs, mediastinal radiation should only be given if oncologic benefit clearly outweighs cardiovascular harm.

Keywords:

• Lymphoma
• cancer survivorship
• cardiovascular disease
• myocardial infarction
• radiation therapy

Background

Lymphoma survivors experience an increased risk of cardiovascular disease that may manifest long past treatment completion [1–7]. Radiotherapy (RT) to the mediastinum, in particular, can damage the heart and surrounding vasculature [8–10]. In survivors of lymphoma, mediastinal RT has been associated with an increased risks of atherosclerosis, cardiac ischemia, myocardial infarction (MI), and cardiovascular mortality [9,11–13].

Traditional cardiovascular risk factors, including hypertension, dyslipidemia, and diabetes mellitus, independently contribute to the risk of MI among lymphoma survivors [2,14]. Among childhood cancer survivors (including those diagnosed with lymphoma), hypertension ascertained in the years after treatment can heighten the cardiovascular risk incurred by radiation [15]. Less is known about adults diagnosed with lymphoma who have a history of cardiovascular risk factors and whether cardiovascular risk factors interact with mediastinal RT to increase risk of MI. Our study aimed to understand whether traditional cardiovascular risk factors and mediastinal RT interact to heighten the risk of later MI in a large, national population of NHL survivors with a full range of cardiovascular risk factors.

Material and methods

Data sources

Using linked population registries to ascertain demographics, hospital discharges, outpatient prescriptions, and lymphoma diagnoses and treatments, we identified people diagnosed with aggressive non-Hodgkin lymphoma (NHL) [16–21]. These lymphomas are generally treated with curative intent; after treatment completion, survivors may be free of symptoms but still have an underlying risk of heart disease that could be mitigated with preventive efforts. We included patients diagnosed at age 15 or older with lymphoma between 2000 and 2010, followed from one year (‘landmark’ analysis) after diagnosis until December 2016, death, or migration from Denmark. We excluded patients who had a cancer other than lymphoma, who had an MI between diagnosis and the landmark, or who emigrated prior to the landmark. Additionally, patients with missing data on receipt of mediastinal RT were excluded.

Outcome ascertainment

The primary outcome was MI after the landmark, diagnosed during inpatient or outpatient hospital visit or as cause of death using International Classification of Disease version 10 (ICD-10) codes for MI (I21, I21.0, I21.1, I21.2, I21.3, I21.4, I21.9, I22, I22.0, I22.1, I22.8, and I22.9). We conservatively opted for this acute cardiac outcome, rather than a more inclusive outcome of any cardiac ischemia, because MI is both clinically important and reliably ascertained in the data.

Preexisting cardiovascular risk factors among lymphoma survivors

Each patient was coded as having 0 or ≥1 cardiovascular risk factor – hypertension, dyslipidemia, and diabetes – diagnosed after 1994 but before diagnosis of lymphoma. Hypertension and dyslipidemia were ascertained using both the National Patient Register and, to improve sensitivity, the National Prescription Register. Hypertension was defined by the prescription of an anti-hypertensive, including calcium channel blocker, diuretics, angiotensin inhibitors, beta blockers, or alpha blockers. Dyslipidemia was defined by the receipt of statins. Diabetes was ascertained in the National Patient Register. We coded each patient as having 0, 1, or >1 risk factor at diagnosis of lymphoma.

Covariates

Receipt of mediastinal RT (yes/no) was determined from the lymphoma registry (LYFO), and RT field was determined algorithmically based on site of treatment when possible. Missing RT data were abstracted from the medical record to the extent possible; because this was a variable of interest, those with missing RT data were excluded from the analysis. We assessed the presence of cardiovascular disease that was diagnosed prior to lymphoma diagnosis and after 1993 using the National Patient Register (see Appendix for ICD-8 and ICD-10 codes). Vascular disease included MI, coronary artery disease, stable and unstable angina, cardiac arrest, stroke, carotid artery disease, and transient ischemic attacks. Intrinsic heart disease included heart failure, pericardial disease, valvular disease, myocardial disease, and cardiomyopathy without clinical heart failure. We included date of birth and sex from the Central Population Register, and date of diagnosis and histology from LYFO.

Analysis

Our primary analysis was conducted using a landmark analysis approach, with follow-up for MI starting 12 months after diagnosis. We chose this landmark a priori to conservatively reflect a time by which most patients will have completed first-line therapy or, in the setting of primary refractory disease in patients eligible for autologous stem cell transplantation, second-line therapy and subsequent consolidation. Survivors were censored at the date of emigration, new primary cancer diagnosis, lymphoma relapse, or death from any cause. Patients with no event were censored on 31 December 2016. Independent variables in the univariable and multivariable Cox regression analyses were selected a priori to capture a parsimonious set of important predictors of MI: history of any of three cardiovascular risk factors at lymphoma diagnosis, age at diagnosis (years), receipt of RT to the mediastinum, sex, stage, and history of vascular or intrinsic heart disease prior to lymphoma diagnosis. The hypothesized interaction effect of mediastinal RT and presence of cardiovascular risk factors on subsequent MI was tested in a univariable Cox regression modeling the association between MI and a four-level variable (± mediastinal RT and ± ≥1 cardiovascular risk factor), in addition to a multivariable model using a likelihood ratio test. To estimate differences in absolute risk of MI based on presence or absence of mediastinal RT and cardiovascular risk factors, we used the absolute 10-year risk of MI in the reference group (i.e., no cardiovascular risk factors and no mediastinal RT) and transformed the hazard ratio to the absolute scale (see supplementary materials).

All statistical analyses were conducted using R software version 3.4.3 (R Core Development Team, Vienna, Austria), including the ‘survival’ package.

Results

Prior to the landmark, 1,288 patients died, 2 emigrated, and 15 developed a new cancer. Eighty people with missing radiation data were excluded. Among the remaining analytic cohort of 3,151 NHL survivors, 32% (N = 1,002) had at least one cardiovascular risk factor (diabetes, hypertension, or dyslipidemia). (Table 1) The median follow-up time among those who did not develop MI was 6.5 years after the landmark. During follow-up of all survivors, 96 survivors had an MI. A model with a four-level variable indicating both mediastinal RT and cardiovascular risk factors showed a significant association of the presence of both mediastinal RT and ≥1 cardiovascular risk factor with later MI (p < 0.001) (Table 2).

Table 1. Patient, disease, and treatment characteristics.

Characteristic
Age at diagnosis, median (IQR)	63	(54, 72)
Sex, N (%)
Female	1419	(45)
Male	1732	(55)
Stage, N (%)
I	915	(29)
II	550	(17)
III	656	(21)
IV	967	(31)
Unknown	63	(2.0)
Diabetes, N (%)	148	(4.7)
Hypertension, N (%)	678	(22)
Dyslipidemia, N (%)	362	(11)
≥1 Cardiovascular risk factor, N (%)	1002	(32)
Vascular disease, N (%)	315	(10)
Intrinsic heart disease, N (%)	148	(4.7)
Received anthracyclines, N (%)
Yes	2759	(88)
No	284	(9.0)
Missing	108	(3.4)
Mediastinal RT, N (%)	269	(85)

N = 3151. IQR: interquartile range.

Table 2. Univariable Cox regression results for associations of mediastinal RT and cardiovascular risk factors with MI.

Mediastinal RT	# CV risk factors	HR (95% CI)	p-Value
No	0	1.00
No	≥1	2.99 (1.94–4.62)	<0.001
Yes	0	0.95 (0.34–2.66)	0.9
Yes	≥1	6.70 (3.25–13.82)	<0.001

RT: radiation therapy; CV: cardiovascular; HR: hazard ratio.

The multivariable model including an interaction term between having any cardiovascular risk factor before lymphoma diagnosis and mediastinal RT indicated that the interaction term was not statistically significant (p = 0.14); therefore, the primary multivariable model did not contain this interaction term. In multivariable analysis, receipt of mediastinal RT was significantly associated with increased hazard of MI (HR: 1.96; 95% CI: 1.09–3.52), as was having at least one cardiovascular risk factor before lymphoma diagnosis (HR: 2.71, 95% CI: 1.77–4.15). (Table 3). Older age at diagnosis (HR: 1.05; 95% CI: 1.03–1.06) and having intrinsic heart disease before lymphoma diagnosis (HR: 2.88; 95% CI: 1.50–5.51) were significantly associated with increased hazard of MI, while being female was associated with reduced hazard of MI (HR: 0.58; 95% CI: 0.38-0.89).

Table 3. Cox regression results for associations with MI.

	Univariable		Multivariable
	HR (95% CI)	p-Value	HR (95% CI)	p-Value
Age at diagnosis (years)	1.05 (1.03–1.06)	<0.001	1.05 (1.03–1.06)	<0.001
Female versus male sex	0.67 (0.44–1.01)	0.058	0.58 (0.38–0.89)	0.013
≥1 Cardiovascular risk factor	3.33 (2.22–4.99)	<0.001	2.71 (1.77–4.15)	<0.001
Vascular disease	3.41 (2.12–5.51)	<0.001	1.30 (0.74–2.26)	0.4
Intrinsic heart disease	4.52 (2.51–8.12)	<0.001	2.88 (1.50–5.51)	0.001
Mediastinal RT	1.54 (0.86–2.76)	0.15	1.96 (1.09–3.52)	0.025

Note: RT: radiation therapy; HR: hazard ratio; CI: confidence interval.

The 10-year risk of MI was 2% among those without mediastinal RT and without cardiovascular risk factors. With this baseline risk, for a person with cardiovascular risk factors, receiving mediastinal RT is associated with a 4.9 percentage-point increase in the risk of an event (i.e., 10-year risk changes from 5.3% to 10.2%), which equates to one additional MI for every 20patients treated with radiation (Table 4).

Table 4. Predicted absolute 10-year risks of MI.

Mediastinal RT	# CV risk factors	Predicted risk of MI (%)
No	0	2.0
No	≥1	5.3
Yes	0	3.9
Yes	≥1	10.2

RT: radiation therapy; CV: cardiovascular; HR: hazard ratio.

Note. Using hazard ratios from the regression model (Table 3), the absolute 10-year risk of MI is shown for individuals with and without mediastinal RT and CV risk factors, based on the finding that those with neither mediastinal RT and no CV risk factors have a 10-year risk of MI of 2%.

Discussion

In a large, registry-based analysis of lymphoma survivors receiving contemporary treatment, we found increased risk of an MI in patients with cardiovascular risk factors and in those undergoing mediastinal RT. Although we did not find evidence of an interaction on the relative scale, the additive effects on the absolute scale are of clinical relevance. The lack of interaction means that radiation is associated with a doubling of risk of MI, whether or not a patient has a cardiovascular risk factor. Because the baseline risk of MI is higher in patients with at least one cardiovascular risk factor, a doubling of risk is of greater clinical significance in this group.

Radiation-related atherosclerosis may increase over time, and most survivors in our study were followed for under ten years. Studies of mediastinal RT and MI in Hodgkin lymphoma typically demonstrate that elevated MI risk is apparent at approximately five to ten years after diagnosis, with increasing rates over time since treatment [22–25]. This suggests that we may have underestimated the relative risk of mediastinal RT, underscoring our conclusion about the effect on patients with CVRFs.

Adults may already have important cardiovascular risk factors when they are diagnosed with lymphoma. Cardiovascular risk was prevalent in our study population; prior to lymphoma diagnosis, 32% had at least one risk factor, 10% had vascular disease, and 4.7% had intrinsic heart disease. Population-based cohorts such as this one are necessary to understand the impact of these important risk factors, as trials may exclude less healthy patients.

Only 77 survivors had at least one cardiovascular risk factor and received mediastinal RT, and power to detect an effect was further limited by the presence of only 9 MI events in this group. Larger cohorts of NHL survivors or cohorts with longer follow-up may reveal that RT administered to patients with prevalent cardiovascular risk factors is more detrimental to the heart. Interactions of cardiovascular risk factors and mediastinal RT have not been found in other studies of survivors treated as adults.

Cardiovascular risk factors are particularly relevant for adults diagnosed with NHL. To prevent MI among survivors, cardiovascular risk factors should be acknowledged in upfront treatment decisions. Similarly, post-treatment monitoring should address prevalent cardiovascular risk. Guidelines from the National Comprehensive Cancer Network for HL survivors recommend annual blood pressure monitoring and aggressive management of cardiovascular risk factors [26]. Guidelines from the European Society for Medical Oncology (ESMO) recommend screening and monitoring along the same lines as other patients with a high risk of MI [27].

Our study lacked information on lifestyle, including tobacco exposure and physical activity, which we partly address by including three well-defined conditions, partly associated with lifestyle. Ascertaining cardiovascular risk factors using prescription data may undercount diagnosed risk factors; to the extent risk factors are untreated, our estimates of the association with MI risk are likely conservative. Our complete case analysis excluded survivors with missing data on chest radiation, which may have introduced bias; however, there is little reason to expect that survivors with missing data were systematically more or less likely to have cardiovascular risk factors or MI. In addition, the relatively short follow-up time only allows for an investigation into the MI risk in the immediate years after end of treatment. The few MI events in this cohort limited the power to test the interaction between RT and cardiovascular risk factors in multivariable analyses. This study is also limited by the few patients receiving RT, and the lack of dose and 3 D volumetric dose estimates to correlate with outcomes. When this cohort of patients underwent radiation therapy (2000-2010), the introduction of advanced imaging (notably positron emission tomography, or PET), 3 D planning techniques and individualized RT volumes (involved-node RT and involved-site RT) reduced the radiation field sizes. Additionally, standard doses were lowered from 40 Gy or more to 30 Gy. As these techniques continue to improve, cohorts of patients receiving lower doses to the heart will likely experience fewer myocardial infarctions resulting from mediastinal RT [28–31]. Currently, however, there are no published studies of population-based cohorts of lymphoma patients who have these data available and that are followed long enough to detect later MI or other cardiovascular events. Despite the absence of detailed information about radiation dose to the heart, our study still found meaningful associations of receipt of mediastinal RT, as well as preexisting cardiovascular risk factors, and later MI across a population-based cohort of lymphoma survivors.

Our population-based registry study investigated cardiovascular risk among a large cohort of NHL survivors, taking into account the risk of MI among a population with wide variation in age, receipt of mediastinal RT, and presence of preexisting cardiovascular risk factors. As treatments improve for lymphoma, survivors are living longer, and reducing cardiovascular risk is critical. There are currently no guidelines for radiation-induced cardiac toxicity in lymphoma survivors. Given that preexisting cardiovascular risk factors add further to the risk of MI, we suggest that routine evaluation and optimal treatment of preexisting cardiovascular disease should be performed before, during, and after receiving radiation therapy. More specifically, for patients with CVRFs who are candidates for mediastinal RT, we recommend careful evaluation of the dose, field, and fractionation to offset the risk of MI. If the increase in cardiovascular risk appears to outweigh the decrease in oncologic risk, mediastinal RT may not be justified.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Access to data in this manuscript are not possible due to GDPR regulations.

Additional information

Funding

This work was supported by the Leukemia & Lymphoma Society as a Clinical Research Scholar Award (T.S.), National Institutes of Health grant under [Grant R03 CA186223]; The Danish Cancer Society; The Danish Council for Independent Research; and the National Institutes of Health under Cancer Center Support [Grant P30 CA008748] (Memorial Sloan Kettering). Dr. De Nully Brown reports personal fees from Novartis, personal fees from Incyte, and personal fees from Takeda, outside the submitted work. Dr. Specht reports personal fees and non-financial support from Takeda, personal fees from MSD, non-financial support from Merck, personal fees from Kyowa Kirin, and grants from Varian, outside the submitted work. Dr. Vickers has stock options in Opko and receives royalties from the sale of a prostate cancer test, the 4Kscore, marketed by Opko. These have no direct impact on the current paper.

Appendix

ICD-8 and ICD-10 codes for pre-existing cardiovascular diseases and risk factors

Table

Variable	Diagnosis	Codes
Vascular disease	Myocardial infarction (ICD-8)	410, 410.09, 410.99, 411, 411.09, 411.99
	Myocardial infarction (ICD-10)	I21, I21.0, I21.1, I21.2, I21.3, I21.4, I21.9, I22, I22.0, I22.1, I22.8, I22.9
	Coronary artery disease (ICD-8)	412, 412.09, 412.99, 414, 414.09, 414.99
	Coronary artery disease (ICD-10)	I25, I25.1, I25.2, I25.3, I25.4, I25.6, I25.8
	Angina (ICD-8)	413, 413.09, 413.99
	Angina (ICD-10)	I20, I20.0, I20.1, I20.8, I20.9, I24, I24.0, I24.8, I24.9
	Cardiac arrest (ICD-10)	I46, I46.0, I46.1, I46.9
	Stroke (ICD-8)	430, 430.08, 430.09, 431, 431.08, 431.09, 433, 433.09, 433.99, 434, 434.09, 434.99, 436.01, 436.9
	Stroke (ICD-10)	G46, G46.0, G46.1, G46.2, G46.3, G46.4, G46.5, G46.6, G46.7, G46.8, I60.0, I60.1, I60.2, I60.3, I60.4, I60.5, I60.6, I60.7, I60.8, I60.9, I61, I61.0, I61.1, I61.2, I61.3, I61.4, I61.5, I61.6, I61.8, I61.9, I63, I63.0, I63.1, I63.2, I63.3, I63.4, I63.5, I63.6, I63.8, I63.9, I64
	Carotid artery disease (ICD-8)	432, 432.01, 432.02, 432.08, 432.09, 432.9, 432.91, 432.92, 432.98, 432.99
	Carotid artery disease (ICD-10)	I65, I65.0, I65.1, I65.2, I65.8, I65.9, I66, I66.0, I66.1, I66.2, I66.3, I66.8, I66.9
	Transient ischemic attack (ICD-8)	435, 435.09, 435.99
	Transient ischemic attack (ICD-10)	G45, G45.0, G45.1, G45.2, G45.3, G45.4, G45.8, G45.9
Intrinsic heart disease	Heart failure (ICD-8)	427.09, 427.1, 427.11, 427.19, 428, 428.99, 782.49
	Heart failure (ICD-10)	I11.0, I13.0, I13.2, I50, I50.0, I50.1, I50.9
	Pericardial disease (ICD-8)	393, 393.01, 393.08, 393.09, 420, 420.08, 420.09, 423, 423.01, 423.02, 423.08, 423.09
	Pericardial disease (ICD-10)	I30, I30.0, I30.8, I30.9, I31, I31.0, I31.1, I31.2, I31.3, I31.8, I31.9, I32
	Valvular disease (ICD-8)	394, 394.9, 394.91, 394.92, 394.98, 394.99, 395, 395.9, 395.91, 395.92, 395.98, 395.99, 396, 396.9, 396.91, 396.92, 396.93, 396.94, 396.98, 396.99, 397, 424, 424.01, 424.02, 424.08, 424.09, 424.1, 424.11, 424.12, 424.18, 424.19, 424.9, 424.91, 424.92, 424.99
	Valvular disease (ICD-10)	I34, I34.0, I34.1, I34.2, I34.8, I34.9, I35, I35.0, I35.1, I35.2, I35.8, I35.9, I36, I36.0, I36.1, I36.2, I36.8, I36.9, I37, I37.0, I37.1, I37.2, I37.8, I37.9
	Myocardial disease (ICD-8)	422, 422.99
	Myocardial disease (ICD-10)	I40, I40.1, I40.8, I40.9, I41, I41.8, I51.4
	Cardiomyopathy (ICD-8)	425, 425.99, 426, 426.08, 426.09, 427, 429, 429.08, 429.09, 426.00
	Cardiomyopathy (ICD-10)	I25.5, I42, I42.0, I42.1, I42.2, I42.3, I42.4, I42.5, I42.7, I42.8, I42.9, I43, I43.8, I51.7