Studies on the aetiology and treatment of post-stroke hand oedema and shoulder-hand syndrome (SHS) published from January 1973 until August 1998 were identified. Eleven studies were included with at least some control for confounding. These were evaluated on 11 methodological criteria and by standardized effect sizes. There were five aetiological studies: four cohort studies and one study consisting of two case series using a within-subjects design. The matters investigated included lymph scintigraphy in hand oedema, bone scintigraphy, putative risk factors and the existence of autonomic dysregulation and peripheral nerve lesions in SHS. There were six therapeutic studies: one randomized controlled trial, one non-randomized controlled trial, one cohort study and three case series, of which two studies used a within-subjects design. These studies investigated continuous passive motion and neuromuscular stimulation in hand oedema as well as oral corticosteroids, intramuscular calcitonin and trauma prevention in SHS. A great diversity of pathophysiological and therapeutic insight was found. Based on systematic analysis of the literature, the following conclusions seem justified: (i) the shoulder is involved in only half of the cases with painful swelling of wrist and hand, suggesting a "wrist-hand syndrome" between simple hand oedema and SHS; (ii) hand oedema is not lymphoedema; (iii) SHS usually coincides with increased arterial blood flow; (iv) trauma causes aseptic joint inflammations in SHS; (v) no specific treatment has yet proven its advantage over other physical methods for reducing hand oedema; and (vi) oral corticosteroids are the most effective treatment for SHS.