Stroke (cerebrovascular accident) is divided into two types according to its underlying pathology. Ischaemic stroke is caused by a blocked arterial blood vessel, whereas haemorrhagic stroke is due to an arterial blood vessel rupture in, or surrounding, the brain (ASA 2005). Several risk factors are common to both types of stroke and include gender, increasing age, family history, race, hypertension, blood disorders, excessive alcohol, and drug abuse. Risk factors for ischaemic stroke are similar to those for ischaemic heart disease and include raised blood cholesterol, smoking, pre‐existing markers of atherosclerosis, and diabetes mellitus (ASA 2005). Recent studies also indicate that the risk of stroke may be higher in women during pregnancy and the 6 weeks following childbirth. For every five deaths from stroke, two occur in men and three in women (ASA 2005).

Stroke is the third leading cause of death after heart disease and cancer, and the first leading cause of serious, long‐term disability in the United States, Canada and Japan (Bonita 1990, ASA 2005). There are about 600,000 new strokes each year in the European Union (EUSI 2007), and nearly 163,000 people die from stroke each a year in USA (ASA 2005). In 2005, the economic burden of stroke‐related medical costs and disability was estimated as 56.8 billion US dollars in United States (AHA 2005). Stroke causes about 10% of all deaths and about 25% of all chronic disability in Australia (NHMRC 1996). In China, the incidence of stroke varies widely from 40 to 3300 per 100,000 in populations aged greater than 40 years, and stroke death rates were 108 to 130 per 100,000, the second highest in the world (Yan 1998). In China there are 1.5 million new strokes occurring each year and about 6 million people currently suffering with stroke (Wang 2004; Wu 1994; Wu 2001). Furthermore, there is evidence of an increasing trend in the age standardised incidence of stroke (Wu 2001).

Prevention is considered to be the best approach to reducing the burden of stroke (Goldstein 2001; Gorelick 1995; NHMRC 1996). The primary prevention strategy for stroke aims to change lifestyles among those who are at high risk of stroke through health education and control of risk factors. Relevant interventions include: regular screening and treatment for hypertension in adults; smoking cessation advice or support for all current smokers; among diabetics careful control of hypertension and glycaemia control to reduce microvascular complications; antithrombotic therapy for patients with nonvalvular atrial fibrillation based on an assessment of their risk of embolism and risk of bleeding complications; weight reduction in overweight persons on the basis of the associated increase in co‐morbid conditions that can lead to stroke; promotion of a healthy diet containing at least five daily servings of fruits and vegetables; no more than two alcoholic drinks per day for men and one drink per day for nonpregnant women. For those with a past history of ischaemic stroke or transient ischaemic attack, antithrombotic treatment with aspirin is indicated (BMJ 2002; NHMRC 1996). The role of aspirin in primary prevention among those at high risk of cardiovascular events is now established but remains controversial among people at intermediate and low risk (Ridker 2005). A Cochrane review showed that anticoagulants are beneficial and without serious adverse effects for people with nonrheumatic atrial fibrillation and recent cerebral ischaemia (Saxena 2004).

Traditional Chinese medicine Chuanxiong is the dried root of the plant Szechuan lovage, Latin botanical name Ligusticum chuanxiong Hort. The main active ingredients of Chuanxiong are ferulic acid and tetramethylpyrazine (ligustrazine) (Tang 1999; Xiao 2002). These two elements have significant effects of inhibiting platelet aggregation, reducing both blood viscosity and haematocrit (Gao 1989; Lin 1991; Tan 2003; Xiao 1994), increasing the level of superoxide dismutase (SOD), reducing malonaldehyde (MDA) and increase nitrogen monoxide (NO) concentration in both hypoxic and ischaemic damaged brain (Huang 2004) and in endothelial cells affected by both hypoxia and reduced glucose (Lin 2004). Ligustrazine also can dilate blood vessels, lower blood pressure, inhibit thrombosis, regulate lipid metabolism, and is a calcium antagonist (Chen 1999), and may prevent and reverse atherosclerosis (Mei 2004).

The available preparations of Chuanxiong compounds have different constituents, but most are based on the formulation "Bu Yang Huan Wu Tang" including Chuanxiong, Huangqi (radix astragali), Danggui (Chinese angelica), Honghua (safflower), Chishao (red paeony root), Dilong (earthworm), and Taoren (peach seed). Chuanxiong may be prepared as a capsule, a water‐based injection, or as decoction (a kind of herb soup). Although the effects of Chuanxiong may vary depending on its constituents and mode of use, some evidence suggests broadly similar effects for all preparations. For example, Nao‐an capsule, used for preventing stroke, consists of Chuanxiong, Danggui (Chinese angelica), Honghua (safflower) and Ren‐shen (ginseng), and has been shown to inhibit thrombosis induced by adenosine diphosphate (ADP) better than aspirin (Chen 1998a), and improve haemodynamic status compared with a control group in patients at high risk of stroke (Wang 2004). It also reduces brain arterial blood vessel resistance (Chen 1998b). Another preparation "Zhong Fang" powder consists of Chuanxiong, Huangqi (radix astagali), Cangzu (Chinese atractylodes rhizome), and Puhuang (cattail pollen), but appears to have similar effects to Nao‐an capsule (Zhou 1999). Chuanxiong and Danggui both contain the active element Ferulic acid (aweishuan) and are often used together in prevention of treatment of cardiovascular diseases. A companion protocol on the use of Chuanxiong compounds for treatment of stroke has been published (Jia 2006).

Although herbal medicine treatment, and its method of manufacture, is widely accepted in China, most of the constituents of the pharmacological preparations used in trials cannot be specified precisely. This is in marked contrast to industrially manufactured pharmacological agents used in Western medicine, in which the chemical constituents, their quantities, and the percentage of any impurities or contaminants, are very precisely known, and the variation between different production batches is kept within specified limits. Variation between formulations and batches of treatments is an inevitable consequence of the nature of Chinese traditional medicine (though the government also specifies the limits of variation that are acceptable). This variation is a factor that may contribute to any heterogeneity between different study results. Furthermore, one must accept that the overall treatment concept for Chinese traditional medicine is different to Western medicine.

While there is some evidence to suggest plausible biological mechanisms whereby Chuanxiong and its compounds might reduce the incidence of stroke in high risk people, the evidence relating to clinical outcomes is currently limited. There is also evidence to indicate that not all herbal medicine is free from harm. There are concerns regarding adverse events, e.g. allergic reactions and Chinese herbal nephropathy (CHN) (Lampert 2002; Lord 2001; Nortier 2000). Despite these concerns, herbal medicines are widely used. In this systematic review we wish to answer the question of whether the available clinical evidence supports the use of Chuanxiong and/or its compounds for preventing stroke.