Most Chinese doctors and patients have, for example, long regarded acupuncture as an effective treatment for stroke, using it to improve motor, speech, and other functions that have been destroyed. One survey showed 66 per cent of Chinese doctors use acupuncture routinely to treat the effects of stroke, with 63 per cent believing it to be effective. Some 36 per cent of Chinese doctors think the effectiveness of acupuncture remains uncertain, perhaps because the scientific basis for it remains so new. Recently, however, systematic scientific studies of acupuncture's effects in such treatment has begun. Almost all trials as a treatment for stroke conducted in China have been positive. But another study in the UK showed research conducted in several countries was uniformly favourable to acupuncture as a treatment for stroke damage. Indeed, all the trials before June 1995 in China, Japan, Hong Kong and Taiwan were deemed positive by the UK researchers.

The Cochrane Collaboration, an international non-profit organisation providing healthcare information, conducted a systematic review of the benefits of acupuncture, including 14 trials, of which 10 were conducted in China, involving 1,208 patients. Acupuncture started within 30 days of stroke onset, with control groups receiving a placebo - sham acupuncture - or no treatment. Compared to patients who received sham acupuncture or no treatment, far fewer of those who received acupuncture died or became invalid within three months. After three months or more of treatment, the numbers rose even more significantly in favour of those treated by acupuncture.

Oddly, when comparing acupuncture with sham acupuncture, the statistical difference concerning death or the need for institutional care was small. Indeed, another analysis, including 14 randomised trials with 1,213 patients six months after a stroke, compared the effects of conventional stroke rehabilitation with and without acupuncture as a supplement. Acupuncture was found to have no additional effect on motor recovery, but a small positive effect on disability, which may be due to a true placebo effect, or to the varied quality of the trials.

Indeed, the efficacy of acupuncture without stroke rehabilitation remains uncertain, mainly because of the poor quality of such studies - a problem that has affected most of the trials undertaken so far. For example, many trials did not describe their method of measuring statistical relevance, and only called themselves "randomised controlled trials". Only four trials, with 373 patients, could provide data on death or dependency (becoming dependent on others for activities of daily living). And only three - all conducted in Europe - used sham acupuncture as a control, whereas no acupuncture trials controlled with placebos or sham acupuncture have been reported in China, owing to high cultural barriers.

Part of the problem in China is choosing the patients to be studied. After all, patients come to hospitals that practice traditional Chinese medicine because they already believe in acupuncture and are likely to be using it to treat another illness. Getting such patients to accept being put into a control group where no acupuncture is applied is almost impossible. This makes random studies of the effects of acupuncture on stroke patients particularly difficult in Chinese hospitals.

More fundamentally, Chinese doctors who use traditional medicine argue that outcome measures of acupuncture as a treatment for stroke should be different from conventional measures used in Western trials, because the theory of acupuncture is different. But no one has yet specified what the appropriate outcome measures for acupuncture should be.

A recent study in China, including 862 patients between the second and tenth day after acute stroke, is the largest truly randomised trial using death or dependency as the primary outcome. The patients were randomly allocated to an acupuncture group receiving acupuncture treatment plus routine treatment, or a control group receiving routine treatment alone. The results could prove useful for therapy of stroke patients.

At present, however, there is insufficient evidence for routine use of acupuncture in stroke. More high-quality trials are needed. We do know that acupuncture is a safe, inexpensive, widely accepted and potentially effective therapy. Particularly at a time when standard rehabilitation is not available for most Chinese stroke patients, few will wait to try acupuncture until its role is confirmed or refuted by reliable evidence.

Ming Liu

Ming Liu is a Professor of Neurology, West China Hospital of Sichuan University, China.

Copyright: Project Syndicate