Risk of Coronary Artery Disease in Patients With
Traumatic Intracranial Hemorrhage
A Nationwide, Population-Based Cohort Study
Wei-Shiang Lin, MD, Chin-Sheng Lin, MD, PhD, Jun-Ting Liou, MD, Wen-Yu Lin, MD,
Cheng-Li Lin, MSc, Shu-Meng Cheng, MD, PhD, I-Ching Lin, MD, PhD, and Chia-Hung Kao, MD
INTRODUCTION
Traumatic intracranial hemorrhage (ICH), a worldwide health problem, is a major cause of mortality in young
adults, and it is associated with high rates of lifelong disabilities among survivors.1,2 In addition to disabilities, survivors of traumatic ICH often experience neurocognitive deficits and psychological problems, such as depression, poor decisionmaking ability, and impulsive aggressive behaviors, leading
to poor community, social, and vocational integration.1 Previous studies have suggested that the detrimental effects of traumatic brain injury (TBI) often extend beyond the hemorrhage area. Metabolic changes have also been observed in
systemic regions distant from the focal hemorrhagic lesions.3,4 Moreover, moderate-to-severe TBI induces several inflammatory signals, which increase proinflammatory cytokine and
chemokine release, resulting in monocyte activation and infiltration, glial activation, neuronal and myelin loss, and persistent
inflammation.5,6 This evidence indicates the systemic injuries and inflammatory reactions following TBI.
Coronary artery disease (CAD), with atherosclerosis being the main pathophysiological mechanism, is the leading cause of morbidity and mortality worldwide.7 Atherosclerosis, a chronic inflammatory disease of the arterial walls, is characterized by the formation of lipid-laden lesions. Furthermore, lipid deposition,
both innate and adaptive inflammation, such as T-cell and macrophage activation, and oxidative stress contribute to
atherosclerosis development.8–10 Among them, inflammation mechanically alters conventional CAD risk factors, such as diabetes mellitus (DM), hyperlipidemia, hypertension, and biological modifications of the arterial walls, which play a major role in atherosclerosis.10
Although inflammatory mechanisms potentially link traumatic ICH and CAD, evidence regarding this association is
limited. By analyzing Taiwan’s National Health Insurance Research Database (NHIRD), we explored the risk of CAD in patients with traumatic ICH and observed an association between them. Therefore, both patients and clinicians should consider the risk of CAD following traumatic ICH.
METHODS Data Source
The National Health Insurance (NHI) program, launched by the Taiwan government in 1995, provides comprehensive health care coverage for 99% of the residents in Taiwan (registered in NHIRD, Taiwan; http://nhird.nhri.org.tw/en/
Background.html). The National Health Research Institute audits and releases the NHIRD for use in health service studies. The NHIRD comprises reimbursement claims data, including registry of beneficiary, medical record for each insured patient, and medical services. Following regulations implemented by the Department of Health, the identity of each patient was encrypted for privacy and data security. For this study, we used a subset of the NHIRD, including files of inpatient claims and Registry of Beneficiaries. This study
was exempted from a complete ethical review by the International Review Board, China Medical University, and Hospital
Research Ethics Committee (IRB permit number:
CMUH104-REC2-115).
Patients
From the inpatient claims data, we identified patients
newly diagnosed with traumatic ICH (International Classification of Diseases, Ninth Revision, Clinical Modification
[ICD-9-CM] 852 and 853) from 2000 to 2010. The initial
traumatic ICH diagnosis admission date was set as the index date. The exclusion criteria are outlined as follows: diagnosis of CAD (ICD-9-CM 410-414) at the baseline, aged <20 years, or having incomplete demographic information. Patients with traumatic ICH were frequency-matched to those without traumatic ICH and CAD at the baseline and by the year of
the index date and age (5-year intervals) by using the aforementioned exclusion criteria. All patients were monitored
until CAD was diagnosed, loss to follow-up, death, withdrawal from the NHI program, or December 31, 2011, whichever occurred first.
Comorbidities
Each patient was screened at the baseline for preexisting comorbidities, including DM (ICD-9-CM 250), hypertension (ICD-9-CM 401-405), hyperlipidemia (ICD-9-CM 272), chronic obstructive pulmonary disease (COPD; ICD-9-CM 490-496), chronic kidney disease (CKD; ICD-9-CM 580-
589), congestive heart failure (CHF; ICD-9-CM 428), depression (ICD-9-CM 296.2, 296.3, 300.4, and 311), and anxiety
(ICD-9-CM 300).
Statistical Analyses
The demographic and clinical characteristics, including
sex, age (_49, 50–64, and _65 years), and baseline comorbidity, of all patients were compared using the x2 test. We used the Student t test for examining the continuous variables; we plotted the Kaplan–Meier curve for measuring the cumulative incidence of CAD for each cohort and tested the difference by using the log-rank test. The overall sex-, age-, and comorbidity-specic incidences of CAD (per 1000 person-year) were calculated for each cohort. Univariate and multivariate Cox proportional
hazard regression models (hazard ratios [HRs] and 95% confidence intervals [CIs]) were used for assessing the risk of CAD
associated with traumatic ICH and then comparing the risk
between the 2 cohorts. The multivariate model was simultaneously adjusted for age, sex, DM, hypertension, hyperlipidemia,
COPD, CKD, CHF, depression, and anxiety. Data were
analyzed for evaluating the CAD-associated risk factors and the coeffects of CAD and traumatic ICH. All analyses were conducted
using SAS statistical software (Version 9.4 forWindows;
SAS Institute, Inc, Cary, NC). A 2-tailed P value of 0.05 was considered significant.
RESULTS
Our study included 128,997 patients with traumatic ICH and 257,994 patients without traumatic ICH. Among patients with traumatic ICH, 66.3% were males, and 40.1% were aged
<49 years (Table 1). The mean age of patients with and without traumatic ICH was 55.3_19.4 and 54.9_13.8 years, respectively.
Furthermore, patients with traumatic ICH were more
predisposed to the comorbidities, such as DM, hypertension, hyperlipidemia, COPD, CKD, CHF, depression, and anxiety, compared with those without traumatic ICH. The mean followup duration was 4.32_3.41 and 5.41_3.16 years in patients
with and without traumatic ICH, respectively. The Kaplan–
Meier analysis results show that the cumulative incidence curves of CAD were significantly higher in the traumatic ICH cohort than in the comparison cohort (log-rank test P<0.001). The overall incidence of CAD in the traumatic ICH cohort was 11.4 per 1000 person-years, which was 1.26- fold higher than that in the comparison cohort (8.97 per 1000 person-years). After adjustment for age, sex, and comorbidity, the risk of CAD was higher in patients with traumatic ICH than in those without traumatic ICH (adjusted HR [aHR]¼1.16, 95% CI¼1.13–1.20). Furthermore, the risk of CAD in patients with traumatic ICH, but without comorbidities, was significantly higher than that in patients without traumatic ICH
stratified by sex (females: aHR¼1.20, 95% CI¼1.27–1.41;
males: aHR¼1.15, 95% CI¼1.11–1.19), age (_49 years:
aHR¼1.44, 95% CI¼1.29–1.60; 50–64 years: aHR¼1.07, 95% CI¼1.00–1.15; >65 years: aHR¼1.09, 95% CI¼1.05–
1.13), and comorbidity (patients without comorbidity:
aHR¼1.22, 95% CI¼1.17–1.28).
Compared with patients without traumatic ICH or underlying comorbidities, those with only traumatic ICH or comorbidities were associated with a higher risk of CAD (Table 3).
Compared with patients with traumatic ICH without comorbidities, those with _6 comorbidities had a significantly increased
risk of CAD (aHR¼15.9, 95% CI¼7.93–31.7), followed by those with 5 (aHR¼6.61, 95% CI¼4.46–9.80), 4
(aHR¼5.95, 95% CI¼5.06–6.98), 3 (aHR¼4.16, 95%
CI¼3.79–4.56), and 2 (aHR¼2.75, 95% CI¼2.59–2.92) comorbidities and those with only 1 comorbidity (aHR¼2.02, 2.02, 95% CI¼1.92–2.12). DISCUSSION
This large-scale, population-based study indicated that patients with traumatic ICH, both men and women, have a significantly higher risk of CAD than those without traumatic ICH, particularly those without comorbidities (Fig. 1, Table 2).
After adjustment for age, sex, and underlying comorbidities, such as DM, hypertension, and hyperlipidemia, patients with
ICH were associated with a 16% increased risk of CAD (Table 2). ICH is an independent risk factor for CAD; patients
with ICH have additive effects on the risk of CAD when combined with other comorbidities (Table 3), suggesting an association between traumatic ICH and CAD.
The NHIRD is a large-scale, nationwide, population-based database, covering approximately 23 million residents in Taiwan and thus permitting several study designs, such as crosssectional, case-control, retrospective cohort analyses, and
family studies.11,12 Because all patients in the NHI program receive optimal medical care at low costs, loss to follow-up is low, which may provide approximately 20 years of follow-up data of the enrolled population. Advantages of the NHIRD
include its large sample size and lack of selection and participation bias.12 In this study, specialists diagnosed and coded (ICD-
9-CM) CAD and traumatic ICH according to the standard diagnosis criteria, including typical symptoms and signs, laboratory data, and imaging findings. Because these data are monitored and strictly evaluated by the Bureau of NHI for reimbursement purposes, the diagnoses of CAD and traumatic ICH are highly reliable.
The underlying mechanisms in the association between traumatic ICH and CAD remain unexplored. Previous studies have suggested that TBI induces interleukin (IL)-1 and tumor necrosis factor-alpha (TNF-a), which triggers an inflammatory cascade and results in an impaired blood–brain barrier and
neuronal and myelin loss.6,13–16 Moreover, TBI activates the innate immune system, resulting in the generation of reactive oxygen species (ROS), such as superoxide or peroxynitrite, which then cause oxidative damage and long-term neurological deficits.
17–20 These inflammatory reactions occur in the central
nervous and systemic circulatory system.21 In atherosclerosis, proinflammatory cytokines, such as IL-1 andTNF-a, play central roles in the inflammatory cascade.22,23 Both in vitro and in vivo investigations suggest that the deposition of cholesterol crystals activates inflammasome, leading to IL-1b production, further
initiating fatty streaks and promoting local atherosclerotic progression.
24,25 Tsimikas et al demonstrated that proinflammatory
IL-1 genotypes were associated with the CAD risk and cardiovascular events, which was mediated by oxidized phospholipids
and lipoprotein(a).22 On the basis of the inflammatory nature of atherosclerosis, trials involving anti-inflammatory therapies targeting IL-1 and TNF-a in cardiovascular disease are ongoing.26
Moreover, the retained low-density lipoproteins undergo oxidation
and modification by ROS and then activate endothelial cells and facilitate foam cell formation, which is considered the
initial step in the formation of atherosclerotic plaque.8,9,27,28 The inflammatory mechanisms linking TBI and atherosclerosis may partly explain the association between traumatic ICHandCADin the present study.
Our study suggests that the incidence of CAD increased in elderly patients and in those with comorbidities, which is
consistent with the observation in previous reports.29–31 However, after adjustment for the potential confounding effects by
conducting a multivariate Cox proportional hazard regression analysis, young age (<49 years) and the absence of comorbidities remained the independent predictors of CAD in patients
with traumatic ICH (Table 2). Younger patients are considered to have a lower incidence of conventional CAD risk factors, such as DM, hypertension, and hyperlipidemia. After stratification for age, aHRs of CAD were the highest in younger patients.
This is possibly because these patients are more predisposed to traumatic ICH alone or the presence and accumulation of complex comorbidities and potential confounding factors
reduce the effects of traumatic ICH.
We evaluated the coeffects for CAD in patients with
traumatic ICH and CAD-related risk factors and observed that although traumatic ICH significantly contributes to the CAD risk, conventional CAD risk factors, such as DM, hypertension, and hyperlipidemia, predominate the development of CAD (Table 3). However, compared with the general population,
patients with traumatic ICH and DM, hypertension, hyperlipidemia, COPD, CKD, and CHF exhibited multiplicative risks of
developing CAD. These interaction analysis results revealed that the potential effects of traumatic ICH on CAD must be considered in clinical practice.
There were some limitations in the study: not available
personal data of the lifestyles or habits as potential confounding factors for CAD in the NHIRD (such as body mass index,
socioeconomic status, family history, smoking, and alcohol);
lack of the individual data of the severity of ICH and CAD, functional status after traumatic ICH, and outcomes with ICH and CAD; the medication compliance could not be evaluated by the NHIRD, leading to the underestimation of the medication effects; we could not use the additional procedure codes to verify the CAD population, although the diagnosis of CAD was strictly monitored by certified medical reimbursement specialists in the NHIRD; because we used ICD-9 codes to evaluate the risk of CAD instead of the screening program, therefore these asymptomatic patients might not ask health care until they had the symptoms or sign of CAD, leading to the underestimation of the risk of asymptomatic CAD in these ICH patients; we could not specify the spectrums of CAD (such as stable and unstable angina and acute myocardial infarction) in this study; and the evidence of the methodological quality is usually lower by a
retrospective cohort study (which lacks these necessary adjustments including these possibly unmeasured or unknown confounding factors) than that by prospective randomized trials.
As per our review of relevant literature, this is the first
study to analyze data from the NHIRD and reveal the association between traumatic ICH and an increased risk of CAD in a
large-scale, population-based cohort. Patients with traumatic
ICH, particularly those with DM, hypertension, and hyperlipidemia, should be educated about the risk of CAD. Although
additional prospective randomized studies are necessary for verifying the effects of traumatic ICH on CAD, meticulous
evaluation and aggressive risk reduction for CAD are recommended in patients with traumatic ICH.
