The transcranial Doppler (TCD) pulsatility index (PI) describes the shape of a spectral waveform and measures cerebral vascular resistance.¹ High PI indicates high-resistance vascular beds with low diastolic flow and peaked waveforms. Low PI indicates low-resistance beds with high diastolic flow and rounded waveforms. High PI suggests increased downstream vascular resistance in the cerebral circulation. PI increases with age and the presence of hypertension, diabetes, vascular dementia, and small artery disease. PI is also influenced by physiological factors, such as the partial pressures of oxygen and carbon dioxide and arterial pressure.² The major factors that affect PI are cerebral blood flow velocity (CBFV) and blood viscosity (BV).³ CBFV is influenced by the length and cross sectional area of the vessel, age, sex, the partial pressures of oxygen and carbon dioxide, and arterial pressure.⁴ BV is the measurement of blood resistance to flow, and is characterized by blood thickness and stickiness.^5-7 Increased BV is related to elevated concentrations of blood cells and plasma components and low deformability and high aggregability of erythrocytes.^7,8

We previously reported that the middle cerebral artery (MCA) PI was significantly associated with BV in acute ischemic stroke patients.⁹ Increased mean MCA PI was associated with old age, presence of diabetes, and increased BV in our study. We also demonstrated that prior antithrombotic use was highly related to decreased BV in acute ischemic stroke patients.¹⁰ Inhibition of platelet aggregation and enhancement of erythrocyte deformability may be associated with decreased BV when antithrombotic drugs are used. Evidence has begun to emerge concerning the relationship between the cytochrome P450 2C19 (CYP2C19) polymorphisms in clopidogrel-treated patients and stroke recurrence and mortality following ischemic stroke,^11,12 but there is no evidence that CYP2C19 polymorphisms in clopidogrel-treated patients influence cerebral vessel viscosity after ischemic stroke.

Considering that TCD PI is associated with BV in acute ischemic stroke patients, we hypothesized that CYP2C19 polymorphisms in clopidogrel-treated patients would influence cerebral vessel viscosity, which was related to TCD PI, and this effect could be demonstrated by serial TCD PI evaluations. Hence, this study was designed to evaluate the difference between the propensities of a good genotype group for clopidogrel metabolism and a poor genotype group to reduce the PI in acute ischemic stroke patients on clopidogrel treatment using serial TCD examinations.

Fig. 1 shows the patient flow and the identified reasons for exclusion from the study. In total, 419 patients were screened during the trial. Of these, 161 patients refused to participate in the study, 123 patients were not eligible for the study, and 51 patients showed poor temporal window. Therefore, 84 patients (20%) were enrolled for this study. Brain diffusion-weighted imaging was performed in 95% of the patients. All patients underwent at least one vascular imaging study (MRA in 96% and CTA in 14%). Anterior circulation stroke was seen in 52% of the patients. The most frequent stroke subtype was lacunar stroke (41, 49%), followed by stroke of undetermined etiology negative work-up (29, 35%) and large artery atherosclerosis (14, 16%). The patient baseline characteristics are shown in Table 1. The mean age was 64.5±12.4 years, and 31% of the patients were women. Of these, 67% had a history of hypertension, 26% had a history of diabetes, 41% had a history of dyslipidemia, 5% had a history of coronary artery disease, and 36% were current smokers. Additionally, 21% of the patients were on antiplatelets, 25% were on statins, and 45% were on antihypertensive drugs at admission. There were no differences in the baseline characteristics among the patients with different stroke subtypes. The relevant artery velocity and PI values were also not different among the groups. PSV, EDV, and mFV were highly associated with PI. PSV was positively correlated with PI (r=0.437, p=0.029). mFV and EDV was negatively related to PI (r<-0.439, p<0.023).

During the trial, aspirin and clopidogrel were given to 77 patients (92%), and clopidogrel alone was given to the remaining 8%. For this study, patients were categorized into decreased-PI and increased-PI groups. Table 2 shows the patient characteristics and laboratory findings according to the PI changes. No significant differences were observed in the medical histories and laboratory findings between the patients of the two groups. Good genotype for clopidogrel metabolism was found in 38% of the decreased-PI patients and 25% of the increased-PI patients (p=0.252). At the end of the 180-day study, we observed that the MCA and BA PI changed significantly with changes in Hct (p<0.013), suggesting that Hct was the major determinant factor of PI. Table 3 shows the correlation analysis of the relevant artery PI with patient characteristics. In the decreased-PI group, 180-day Hb and Hct were negatively associated with PI changes. In the increased-PI group, baseline Hb and Hct were significantly related to PI changes, also suggesting that Hct was the major determinant factor of PI. CYP2C19 genotype status for clopidogrel metabolism was not associated with PI changes in both groups.

TCD PI is the reflection of multifactorial, pleiotropic events that occur in the systemic and cerebral vascular systems. Instructive studies on TCD PI and acute ischemic stroke are limited,^3,15 and no studies have addressed the serial PI changes associated with CYP2C19 polymorphisms in acute ischemic stroke patients on clopidogrel treatment. We speculated that CYP2C19 polymorphisms in clopidogrel-treated patients might influence cerebral vessel viscosity, which was related to TCD PI. However, CYP2C19 genotype status for clopidogrel metabolism was not associated with PI changes in this study. Our previous study demonstrated that in the clopidogrel treatment group, a good CYP2C19 genotype for clopidogrel metabolism was associated with a 31% decrease in the relative risk of recurrent stroke. However, a 41% decrease in the relative risk of recurrent ischemic stroke in the good CYP2C19 genotype-group has been observed.¹⁶ Informative studies on pharmacological therapies for reducing BV, which may be related to decreased PI, are limited.^17,18 One study showed that in patients with subclinical carotid or femoral atherosclerosis, clopidogrel reduced the mean lowshear BV by 18% after 3 weeks of treatment.¹⁸ Aspirin and dipyridamole are more effective in reducing lowshear BV than the efficacy exhibited by aspirin alone.¹⁹ Since our study failed to demonstrate the relationship between the TCD PI changes and the CYP2C19 genotype in clopidogrel-treated patients, further studies that focus on the CYP2C19 genotype and BV are essential to evaluate the effects of antithrombotic therapy on PI changes.

Our study confirmed that Hct was clearly related to PI changes. These results were in concordance with those of previous studies that showed positive correlation between Hct and PI.^4,20,21

The major factors that affect PI are CBFV and BV.³ Decreases in Hct are associated with increased CBFV, which, in turn, is related to decreased PI.^20,21 BV reflects frictional interactions between blood components and erythrocytes within the systemic vascular system.¹⁰ The major determinants of BV are erythrocyte aggregation and deformability, Hct, and plasma viscosity. Of these, Hct is the most essential factor, and alterations in Hct contribute to rheological discrepancies.²² Increases in Hct are related to increased BV. The positive association of Hct to BV and arterial oxygen carrying capacity may be a plausible explanation for this phenomenon.

The strength of this study was that it incorporated a prospective longitudinal design using serial TCD examinations. Unlike traditional cross-sectional single studies, this study was used to evaluate the Hct and TCD PI changes in acute ischemic stroke patients. In the decreased-PI group, 180-day Hct was negatively associated with PI changes. Baseline Hct was positively related to PI changes in the increased PI group. However, there were several limitations, including small sample size, in this study. Of the 419 patients that were screened during the trial, only 84 (20%) were enrolled for this study. Though a previous report indicated that 41% of the Korean participants belonged to the good genotype-group for clopidogrel metabolism,²³ our study showed that only 33% belonged to the good genotype-group. These results might be due to the small sample size, which might have affected the study results. Second, clopidogrel response might have been affected by several clinical factors. Owing to the small sample size, we could not evaluate the clinical factors associated with clopidogrel response variability. Finally, hydration might be associated with Hct and BV changes. One study demonstrated that BV was significantly higher at admission, but was normalized after 2 weeks of hydration.⁵ In our study, all blood samples and TCD studies were not performed on the same day at baseline. These limitations should be considered during the interpretation of our data.

In conclusion, the CYP2C19 genotype status was not associated with TCD PI changes in acute ischemic stroke patients on clopidogrel treatment. Hct was the major determinant factor of PI. Physicians should be aware that TCD PI may be affected by Hct in serial TCD examinations. The negative association of Hct with CBFV and its positive association with BV are plausible explanations for this phenomenon.