Association between hypertensive pregnancy disorders and future risk of stroke in Taiwan: a Nationwide population-based retrospective case-control study

Background The incidence of female stroke has increased gradually and has begun occurring at a younger age in recent years. Given that women live longer than men, stroke would cause more negative and longer-term impacts on the rest of the lives of women. There are few related studies on Asian women. We aimed to evaluate stroke risk in Asian women following hypertensive pregnancy disorders. Methods Using the Taiwan National Health Insurance database, we designed a retrospective study that included pregnant women between 2000 and 2013. We selected an age-matched control group of women without hypertensive pregnancy disorders at a 1:3 ratio. The endpoint was any episode of stroke; otherwise, the patients were tracked until December 31, 2013. After the index date until the end of 2013, Cox proportional hazards analysis was used to compare the risk of incident stroke. The risk factors for stroke were determined using Cox proportional regression to calculate the hazard ratio (HR) compared with the control group. Results During the follow-up period, the Kaplan-Meier analysis indicated that patients with hypertensive pregnancy disorders had a significantly higher risk of developing stroke than did patients without hypertensive pregnancy disorders (log-rank test P < 0.001). Multivariate Cox regression analysis demonstrated that the case group had a 2.134-fold increased risk of stroke (HR = 2.134; 95% CI = 1.817–2.505; P < 0.001). Conclusion Our study provided evidence of an increased risk of stroke in patients with hypertensive pregnancy disorders. Compared with those without such disorders, the patients who had experienced the disorders had a 2.134-fold (P < 0.001) higher risk of developing stroke in the future.


Background
Hypertensive pregnancy disorders are a set of pregnancyspecific systemic diseases with a unique pathophysiology [1][2][3][4][5]. There are three main types of the disease: gestational hypertension, pre-eclampsia-eclampsia, and superimposed pre-eclampsia. Preeclampsia is a multisystem disorder accompanied by the new onset of hypertension and end-organ dysfunction with or without proteinuria. The critical conditions of hypertensive pregnancy disorders include eclampsia, hemolysis, elevated liver enzymes, low platelet counts, and disseminated intravascular coagulopathy [6][7][8][9]. Hypertensive pregnancy disorders are critical risk factors for pregnancy-associated stroke [10]. They affect approximately 5% of pregnant women and increase the risk of pregnancy-associated cerebrovascular disease during the intrapartum and postpartum periods [11][12][13][14]. Cerebrovascular dysfunction related to hypertensive pregnancy disorders can lead to stroke, cerebral edema, seizures, and maternal mortality [10,15,16].
A recent study revealed that although women were more likely to survive after stroke, poorer recovery was noted [17]. However, most women did not know they were facing the risk of stroke. According to the JAMA guidelines, an estimated 425,325 new or recurrent strokes occur in women in the United States each year, representing approximately 53.5% of the stroke population. In 2010, women accounted for approximately 60% of stroke-related deaths (77,109 of 129,476 deaths). Approximately 3.8 million women live after having a stroke in the United States [18]. Stroke is also the third-leading cause of death in Taiwan, and the mortality rate of stroke was 7.2% in 2012 [19]. Based on the 1994 National Health Interview Survey, women accounted for approximately 46% of the stroke population in Taiwan [20]. The related risk factors for stroke included hypertension, diabetes, hyperlipidemia, obesity, atrial fibrillation, and smoking [21][22][23][24][25].
However, few studies have mentioned the association between stroke not occurring during pregnancy or the postpartum period and hypertensive pregnancy disorders [26][27][28][29]. Thus, our study determined the risk of future stroke in women in Taiwan with hypertensive pregnancy disorders.

Data sources
The National Health Insurance program was implemented in 1995 in Taiwan, and the National Health Insurance Research Database (NHIRD) records the medical information of all insured people. The database includes approximately 23.74 million people in Taiwan, and it obtained a coverage rate of approximately 99.6% in 2009 [30,31]. For study purposes, from 1 million individuals randomly selected as subjects obtained from the NHIRD, we created a small database. The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) was used for the diagnostic and treatment codes in the NHIRD. With the permission of the National Health Research Institute, our study was able to use the databank. This study was granted approval by the Institutional Review Board of Tri-Service General Hospital (TSGHIRB No. 1-105-05-142).

Study design and sampled participants
This study was conducted with a retrospective casecontrol design with outpatient and inpatient data. We collected our data from January 1, 2000 to December 31, 2013. Among the 989,753 individuals, 43,675 individuals were diagnosed with hypertensive pregnancy disorders (ICD-9 codes 642.0 to 642.9) prior to the index date. The patients were excluded if they met one of the following criteria: diagnosed with hypertensive pregnancy disorders before the index date, diagnosed with one stroke episode before tracking, aged < 12 years, and male sex. At the end of the study, the case group comprised 41,870 individuals. A control group matched by index day and age with propensity scores was enrolled at a 3:1 ratio. The same exclusion criteria as that used for the case group was implemented, and the controls did not have hypertensive pregnancy disorder episodes in the study period; 125,610 individuals comprised the control group. The tracking event refers to the occurrence of stroke, and tracking continued until December 31, 2013. Individuals with any diagnoses of cerebrovascular disease (ICD-9 codes: 436.0 to 436.9 and 437.0 to 437.9) were defined as being diagnosed with stroke ( Fig. 1).

Outcome measures
All study participants were followed from the index date until the onset of stroke (ICD-9 codes: 436.0 to 436.9 and 437.0 to 437.9), withdrawal from the insurance program, or the end of 2013. We eliminated peripartum strokes and postpartum strokes within 3 months (Table S3).

Statistical analysis
We conducted the analyses using SPSS 20 software (SPSS, Inc., Chicago, IL, USA). Descriptive statistics were used for basic information, including percentages, average values, and standard deviations.
Differences in the distribution of age, insurance premium, sex, season, location, urbanization level, comorbidities, and level of hospital between the two groups and between subjects with and without stroke were compared using the chi-square test. The Cox proportional hazards regression model was also applied to assess the influence of hypertensive pregnancy disorders on the risk of stroke. The relevant hazard ratios (HRs) and 95% confidence intervals (CIs) are presented. The difference in stroke risk between the two groups was estimated by the Kaplan-Meier method with the log-rank test. All results were statistically significant if the twotailed p value was less than 0.05.

Results
This study examined 41,870 patients with hypertensive pregnancy disorders and 125,610 controls. At the 13year follow-up, the cumulative incidence of stroke was 0.47% (198/41,870 individuals) in patients with hypertensive pregnancy disorders and 0.56% (698/125,610 individuals) in patients without hypertensive pregnancy disorders ( Fig. 1 and Table S1). The Kaplan-Meier analysis indicated that patients with hypertensive pregnancy disorders had a significantly higher risk of developing stroke than did patients without hypertensive pregnancy disorders (log-rank test P < 0.001) (Fig. 2). Strokes that occurred during the pregnancy and within 3 months postpartum were eliminated. These strokes were only a small portion of the total burden of disease (Table S3).
In the 13th year of follow-up, the case group was more likely to develop stroke than the control group (p < 0.001). At the end of follow-up, compared to the controls, patients with hypertensive pregnancy disorders tended to be younger in age (35.49 years old vs. 36.71 years old; p < 0.001), have lower insurance premiums (95.48% vs. 94.77%; p < 0.001) and have higher rates of Compared to the controls, more patients with hypertensive pregnancy disorders were treated in the hospital center (32.72% vs. 26.89%) and regional hospital (37.65% vs. 32.53%) (P < 0.001) ( Table 1).
The Cox regression analysis of the factors associated with the risk of stroke showed that the crude HR was 2.704 (95% CI = 2.326-3.144, P < 0.001). After adjusting for the season, the urbanization level of residence and monthly income, the adjusted HR was 2.134 (95% CI = 1.817-2.505, P < 0.001). Compared with younger women (aged 12-19 years old), for women between 20 and 29 and 30-39 years old, the adjusted HRs were 1.306 (P < 0.001) and 1.215 (P = 0.011), respectively. Our results indicated that patients with hypertensive pregnancy disorders had a 2.134-fold higher risk of developing stroke than did individuals without hypertensive pregnancy disorders (Table 2 and S2). Those with diabetes mellitus (DM) (P < 0.001), hypertension (HTN) (P < 0.001), hyperlipidemia (P < 0.001), heart disease (P < 0.001), and chronic kidney disease (CKD) (P < 0.001) had a higher risk of developing stroke than those without these comorbidities. A higher incidence of stroke development was observed among hypertensive pregnancy disorder patients who visited the hospital center than among those who visited the local hospital ( Table 2).
The incidence and HR of stroke in populations with or without hypertensive pregnancy disorders relative to those in controls are listed in Table 3. After adjusting for all the other variables, regardless of the other factors, compared with patients without hypertensive pregnancy disorders, those with hypertensive pregnancy disorders had a rate of 1.21 per 1000 patient-years vs. 1.06 per 1000 patient-years in those without gestational hypertension and an HR of stroke that was 2.134-fold (P < 0.001) higher than that of those without gestational hypertension (Table 3 and Fig. 3).

Discussion
Hypertensive pregnancy disorders are the most common cause of hypertension in pregnant women [5]. They are   [32]. Women with chronic high blood pressure experience more complications during pregnancy [12]. High blood pressure may also lead to other long-term health problems after pregnancy. Stroke is an undesirable condition among younger women, as it is considered to have adverse effects on personal status, families and society. Health issues not only affect women but also demand major changes of the whole family. The incidences of both early-and late-onset preeclampsia increased in Taiwan from 2001 to 2014, especially for early-onset disease [33]. Preeclampsia has become an important issue to be aware of. Studies show that hypertensive pregnancy disorders not only affect the pregnancy period but also increase the risk of cardiovascular disease later in life [34,35]. Pregnancy-related hypertension is thought to be an important risk factor for both cerebrovascular disease and intracranial venous thrombosis in some studies [36]. One study based on the general population during the intrapartum and postpartum periods in Taiwan revealed that preeclampsia/eclampsia are the two most common causes of intracranial hemorrhage and the three most common causes of cerebral infarction during the intrapartum and postpartum periods [37].
In one cohort study in Sweden, the authors found that an increased risk of cardiovascular disease and cerebrovascular disease after hypertensive disease of pregnancy persisted in the older population [21]. One study in Taiwan previously revealed that the respective adjusted relative risk of hemorrhagic and ischemic cerebrovascular disease after preeclampsia-eclampsia was much higher, regardless of whether it was within the first 3 days postpartum or within 1 year of the postpartum period [38]. One retrospective cohort study based on the Utah Population Database also showed that women with hypertensive disease of pregnancy have increased mortality risk, particularly with respect to ischemic heart disease and cerebrovascular disease [39].
Our study revealed an important issue in hypertensive pregnancy disorders during pregnancy. Strokes that did not occur in the pregnancy and postpartum timeframes accounted for a main portion of the total burden of disease. Women with hypertensive disorders of pregnancy are at significant risk for stroke at a relatively young age. Interval hypertension control and follow-up after pregnancy is likely to be important. Patient education on future stroke issues, exercise and diet counseling could be involved in the postpartum period. Women with a history of hypertensive pregnancy disorders should be candidates for risk modification, including HTN control, lipid management, and weight management.
Our study included a well-established dataset using population-based research with a large sample size, and we explored hypertensive pregnancy disorders as a risk factor for developing stroke. Nevertheless, there are still some limitations of this study. First, although the coding of the NHIRD has not been validated in the recording of all diseases, there were no reports regarding the coding Season of the severity of stroke or the location of stroke. Data on blood pressure during pregnancy and the control status of hypertensive pregnancy disorders were also unavailable in the database. For these reasons, the effect of hypertensive pregnancy disorders on the severity of stroke or the site of stroke could not be analyzed in detail. Second, the NHIRD registry cannot provide detailed information regarding patients' laboratory results, lifestyles, past history or family history, and we believe some of these factors, such as smoking, may increase the risk of stroke. Third, given that the population comprises younger females, there might be an underlying disease of the brain that was not noted before, and although the population may be small, we believed that it could also contribute to the increased incidence of stroke [40]. Fourth, there are also some factors that may contribute to stroke, such as autoimmune diseases [41][42][43], thyroid disease [43][44][45], oral contraceptive use [46][47][48], and infectious diseases [49][50][51], none of which were discussed in this study.

Conclusions
Our study provided evidence of the increased risk of stroke in patients with hypertensive pregnancy disorders. Compared with those without hypertensive pregnancy disorders, the patients who had experienced such disorders had a 2.134-fold (P < 0.001) higher risk of developing stroke in the future. The occurrence of stroke in younger women is an important issue, especially among those who are in the childbearing stage. The results from this study will provide physicians with strong proof of the need for the cautious treatment of patients with hypertensive pregnancy disorders and awareness of possible future stroke problems.