Abstract
Background and objective
Ischemic stroke (IS) is the 2nd killer in Taiwan. There was no studies of air pollution indicators (API's) in nationwide database, on their relationship to IS recurrence. We aimed to analyze the IS incidences, its recurrence risk related to API's.
Methods
We used LHID 2005 (Longitudinal Health Insurance Database 2005) from National Health Institute Research Database (NHIRD), i.e. the entire original claim data of randomly sampled 1,000,000 beneficiaries in 2005. The national API surveillance database (PM2.5, PM10, CO, SO2, NO,NO2, NOx, O3) was incorporated with it. The IS incidences was computed by obtaining 1st admission for IS (ICD9–433, 434, 435), and the recurrence was on second IS within 30 days of 1st one. To minimize the confounding effect of comorbidities, we used a time-stratified case-crossover study design, proposed by Maclure (1991), for transient effects on acute events' risk; it is characterized by that each subject serves as his or her own control according to fixed individual characteristics, such as age, gender, lifestyle, socio-economic status, genetics and physiological status, etc. The API exposure was analyzed on the re-admission date (lag0), and 1 to 6 day before re-admission (lag1, lag2 ... lag6 etc.).
Results
The IS recurrence-readmission is strongly associated with SO2 peak level 5 days before (lag5 – 95% CI: 1.011–1.023, p=0.004) for patients above 20. Significant association of IS recurrence with PM2.5, SO2 O3, SO2 is noted in different patient groups.
Table 1. Basic demographics of ischemic stroke in Taiwan (2005–2013) Ischemic stroke N % Age, mean ± SD 68.94±12.8 Male 11832 58% Other forms of chronic ischemic heart disease (ICD9–414) 1496 7% Diabetes mellitus (ICD9–250) 6679 32% Essential hypertension (ICD9–401) 9038 44% Disorders of lipoid metabolism (ICD9–272) 4142 20% Heart failure (ICD9–428) 670 3% Hypertensive heart disease (ICD9–402) 2333 11%
Figure 1. Hazard Ratio of specific API's in IS (only shown with statistical significance): those left to line 1.0 with protective effect and vice versa (20–45, 46–64 and 65 above representative meaning for age range).
Conclusion
Specific API's are strongly associated with IS recurrence. We need a prospective cohort study designed on this topic.