Risks of subsequent epilepsy among patients with hypertensive encephalopathy:
A nationwide population-based study
Tzu-Tsao Chung a,1, Chi-Yu Linb,1,Wen-Yen Huang c,d, Cheng-Li Lin e,f, Fung-Chang Sung g,h,i,j, Chia-Hung Kao d,e,⁎
1. Introduction
In 1928, Oppenheimer et al. frst introduced the term “hypertensive encephalopathy” (HE) to describe acute episodes of several cerebral phenomena correlated with hypertension [1]. It is now known as an acute organic brain syndrome characterized by unspecifc neurological symptoms such as headache, visual disturbance, altered mental status, and seizure [2]. The term “reversible posterior leukoencephalopathy syndrome” has also been used because the most common abnormality associated with the syndrome in computed tomography and magnetic resonance images is edema involving white matter in the parietooccipital areas [2–4]. It is suggested that extreme elevation of systemic
blood pressure causes the breakdown of the autoregulatory capability of the brain vasculature, resulting in associated neurological syndromes [4]. Although HE is usually reversible, failure to promptly treat the
dramatic rise in blood pressure may lead to fatal consequences. It is an emergent syndrome that requires correct identifcation and early management.
The exact incidence of HE is unknown.However, hypertensive
crises represent more than one-fourth of all medical emergencies and can result in acute end-organ injuries such as cerebral infarction, acute myocardial infarction, heart failure, acute renal failure, and HE [5].
Epilepsy is commonly concomitant with acute episodes of HE, most likely because of the irritation caused by transudate in the interstitium.
However, the probability and frequency of epilepsy in the durable follow-up of patients with a history of acute episodes of HE remain unclear. The understanding of this issuemay provide information valuable
to follow-up strategies for patients with HE. In this study, we investigated whether the diagnosis of HE is linked
to an increased risk of developing subsequent epilepsy by using the Taiwanese National Health Insurance Research Database (NHIRD). The database is available to researchers in Taiwan and has been extensively used in epidemiologic studies [6]. The wide coverage of this large, nationwide database allowed us to examine the relationship between
HE and the subsequent development of epilepsy.
2. Materials andmethods 2.1. Data sources
This retrospective study used data retrieved fromseveral claims fles of the NHIRD,managed by Taiwan's National Health Insurance Research Institute (NHRI) at the Department of Health. The universal National Health Insurance (NHI) program was implemented in March 1995 in Taiwan and covered approximately 99% of the total 23.74-million residents in 2009. This study analyzed the nationwide populationbased database released by the NHRI from 1996 to 2010 for academic and administrative use. The NHI database includes information on the basic patient demographic status, medical institutions, details of inpatient orders, ambulatory care, expenditures for care, and physicians
providing services. To protect patient privacy, all patient-level information can be retrieved and linked only through scrambled personal identifcation.
This study was approved by the Ethics Review Board of China Medical University (CMU-REC-101-012).
2.2. Study patients
We identifed patients with a diagnosis of hypertension (ICD-9-CM codes 401–405) fromclaims data for inpatients for 1997–2010. Patients aged 20 years and older with newly diagnosed HE (ICD-9-CM code 437.2) were selected for the study cohort. The inpatient diagnosis date was defned as the index date. The comparison cohortwas randomly selected fromthe rest of the hypertensive patientswithout a history ofHE.
For each patient in the study cohort, 4 comparisons were randomly selected, frequency-matched by sex, age (every 5-year span), and the year
of the index date. Patients with any record of epilepsy and/or stroke (ICD-9-CM codes 430–438) before the index date were excluded. In order to exclude acute provoked seizure, we set a lag time of 1 week to exclude those who had seizures in the frst week after the diagnosis of HE. Thus, 5 patients were excluded: 3 in the study cohort and 2 in the comparison cohort.
2.3. Outcome measures
Both cohorts were followed from the index date to the date the patients received the diagnosis of epilepsy (ICD-9-CM code 345) or until
the patients were censored because of lack of follow-up, death, withdrawal from the NHI system, or the end of 2010. Because epilepsy was
an emergent and critical condition clinically, we believed that physicians
made the diagnosis with caution. We supposed that one coding
of 345 was enough to defne the epilepsy. The comorbidities considered in this study included head injury (ICD-9-CM codes 850–854, 959.01), meningitis (ICD-9-CM codes 0130, 0360,0470, 0471, 0478, 0479, 0490, 0491, 0530, 0721, 0942, 1142, 320, 321, 322, 00321, 05472, 09042, 09181, 09882, 10081, 11283, 11501, 11591), encephalitis (ICD-9-CM codes 0136, 0361, 0462, 0520, 0550, 0722, 1390, V050, 062, 063, 064, 323, 09041, 09481), multiple sclerosis (ICD-9-CM code 340), and alcoholism (ICD-9-CM codes 303, 305.00, 305.01, 305.02, 305.03, V11.3).
2.4. Statistical analysis
The sociodemographic distribution and prevalence of comorbidities were compared between the study and comparison cohorts and examined
using theχ2 test. The sex-, age-, and comorbidity-specifc incidence densities of epilepsy were measured and compared for both cohorts.
Poisson regression was used to estimate an incidence rate ratio (IRR) and a 95% confdence interval between the study cohort and the comparison cohort. Multivariate Cox proportional-hazards regression analysis was used to estimate the risk of epilepsy in association with HE, controlling for sociodemographic factors and comorbidities.
The follow-up years were partitioned into 4 segments (≤3 years, 3–6 years, 6–9 years, and N9 years) to observe the change in epilepsy hazard. The cumulative incidence of epilepsy for both the study and the comparison cohorts was calculated using the Kaplan–Meier method, and the difference was tested using the log-rank test. The analyses were performed using the SAS statistical package (version 9.2; SAS Institute Inc., Cary, NC, USA), and the Kaplan–Meier survival curve was plotted using R software (R Foundation for Statistical Computing, Vienna, Austria). The statistical signifcance was accepted at an α-value of 0.05.
3. Results
Among patientswith hypertension but free from stroke,we identifed 5766 patients with HE for the study cohort and selected 23,074 patients for the comparison cohort. Patients in both cohorts were predominantly female and over 65 years of age. Comorbidities were more prevalent in the study cohort than in the comparison cohort, particularly for head injury, meningitis, encephalitis, and alcoholism (Table 1).
The incidence of epilepsy in the study cohort was 2.26-fold greater than that in the comparison cohort (4.17 vs. 1.85 per 1000 personyears),
with an adjusted HR of 2.06 (95% CI: 1.66–2.56) (Table 2). Men were at higher risk than women to have epilepsy in both cohorts. The HE to non-HE adjusted HR was also higher for males than for females (2.27 vs. 1.84). Age-specifc data showed that epilepsy incidence was the highest in patients with HE 20–39 years of age, although there were more cases of epilepsy in the older groups. The study cohort to comparison cohort relative risk of epilepsy was the highest for those aged 40–64 years old, with an adjusted HR of 2.59 (p b 0.001).
Table 2 also shows that epilepsy incidence increased for those with comorbidity,
with the highest incidence for those with meningitis,
followed by those with encephalitis, alcoholism, and head injury.
The incidence of epilepsywas consistently higher in the study cohort with HE than in the comparison cohort during the follow-up period (Table 3). The incidence of epilepsy was the highest during the initial 3 years after HE diagnosis with an adjusted HR of 3.03 compared with
the corresponding comparisons. The cumulative incidence curve for epilepsy showed that the study cohort had a signifcantly higher risk of
epilepsy than the comparison cohort (log-rank test b 0.0001) (Fig. 1).
4. Discussion
To the best of our knowledge, this study is the frst attempt to investigate the risk of epilepsy among patients with HE after adjusting for
patient comorbid medical disorders by using a nationwide populationbased data set. Our study shows that the likelihood of epilepsy development is 2.06-fold greater among patients with HE than among hypertensive patients without encephalopathy. Furthermore, we found that
patients with HE with comorbidities of head injury, meningitis, and alcoholism were at an additionally higher risk of epilepsy than those without the comorbidities.
The overall demographic-specifc incidence of epilepsy after stratifed analyses of sex and age is signifcantly higher in the cohort with HE than in the comparison cohort. We found that the relative risk of epilepsy was higher in men than in women (adjusted HR = 2.27 vs.
1.84), concurring with well-established databases of sex comparisons in epilepsy [7].
Moreover, the relative incidence of epilepsy is still higher in the long-term follow-up, further confrming that the risk of epilepsy in patients with HE is “truly” increased. Furthermore, approximately half
of epilepsy diagnoses occurred within the initial 3 years after the frst episode of HE, with a decreased
relative incidence of epilepsy in the
prolonged follow-up. In a recent meta-analysis, the estimated median incidence of epilepsy was 0.504 per 1000 person-years [8]. However, the overall incidence of epilepsy in our study was high, with a rate of 4.17 per 1000 person-years in the cohort with HE and 1.85 per 1000 person-years in the comparison cohort. This is probably because the participants were relatively older; 62.4% of both cohorts were age 65 or older. However, the age-specifc incidence of epilepsy was the lowest for those aged 40–64, indicating a U-shape association with age. With the highest incidence of epilepsy, younger patients with HE deserve greater attention after the diagnosis of HE.
Changes in the degree of hypertension can generate the autoregulatory dysfunction of cerebral blood fow [9]. Recent reports assert
that HE occurs in 15%–20% of patients inwhommalignant hypertension was present [10,11]. It is an acute organic brain syndrome resulting from disrupted autoregulation of cerebral blood fow. An acute provoked seizure is a frequently concomitant with HE [2,12]. There may
be generalized, focal, or focal with secondarily generalized tonic–clonic convulsions. The pathogenesis of acute provoked seizure following HE is incompletely understood. It seems to be the irritative effects of the fuid in the brain interstitium related to cytotoxic edema or vasogenic edema [1,3,13–15]. The cytotoxic edema results from the infarction caused by thrombosis of arterioles and fbrinoid necrosis [13–15].
Conversely, vasogenic edema is related to hypertensive cerebrovascular endothelial dysfunction or disruption of the blood–brain barrier with increased permeability [3,9,15]. However, the disruption is the most widely accepted basis for HE [16]. Therefore, to reduce acute provoked seizure, neuroprotection following diagnosis of HE deserves further research.
The pathogenesis of later spontaneous unprovoked seizures, the endpoint of the present study, is even less studied and discussed. Itmay refect more permanent structural and physiologic changes within the brain.We found an increased risk of later epilepsy in patients with HE.
This may provide the basis of ongoing research on the pathogenesis and prevention strategy of later spontaneous unprovoked seizures in patients with HE.
The timing of the most development of spontaneous unprovoked
seizures after HE is important.We found that most of the seizures happened in the frst 3 years even though the increased incidence of epilepsy
was found within 6 years after occurence of HE. This pattern is similar to that of epilepsy after a traumatic brain injury. Approximately
40% of individuals with epilepsy after head trauma have onset within 6 months, 50% within 1 year, and 80% within 2 years [17,18]. The
more severe the head injury, the longer the patient is at risk for late seizures.
This information has implications in the follow-up strategy and management of patients with HE.
A particular strength of this study is the use of a nationwide
population-based data set that provides a suffcient sample size and statistical power to explore the link between HE and epilepsy. In addition,
the patients in our study displayed awide range of demographic characteristics, which allowed us to perform stratifed analyses according to
sex, age, and comorbidities. Nevertheless, some insuffciencies in our
study should be addressed. First, additional theoretically relevant confounding variables such as smoking, diabetes, and a family history of
epilepsy could not be included in our analysis because theywere not included in our data set. Further study is needed to clarify the effects
of these factors. Second, we might not be able to completely exclude study subject “misclassifcation”. A patient with HE, with symptoms of headache, visual disturbance, and altered mental status but no seizure, might not seek medical advice and may thus be misclassifed as having hypertension only and be included in the comparison cohort. We believed that this probability was extremely low because few patients
would tolerate acute HE symptoms without medical intervention.
Moreover, our patients could seek medical advice easily because of the high accessibility of medical services in Taiwan.
5. Conclusion
We found that the risk for epilepsy in Taiwan was approximately 2.24-fold greater among patients who had been previously diagnosed with HE compared to those who had not and that this association was entirely independent of age, sex, head injury, meningitis, encephalitis, alcoholism, and multiple sclerosis. Thus, physicians should be aware of HE's link to epilepsy when assessing patients with HE. Furthermore, because approximately half of the epilepsy diagnoses occurred within 3 years from the onset of HE, routine follow-up examinations and control of blood pressure should be performed for at least 3 years.
