Background and Purpose—The reported prevalence of malnutrition after stroke varies widely, whereas it remains unclear which of the estimates is most accurate. The aim of this review was to explore possible sources of this heterogeneity among studies and to evaluate whether the nutritional assessment techniques used were valid.

Methods—A literature search was conducted to identify all studies in which the nutritional state of patients was assessed after inpatient admission for stroke. The percentages of patients identified as malnourished in each study and method of nutritional assessment are reported. For the purposes of this study, an assessment technique was considered valid if at least one form of validity had been demonstrated previously through psychometric evaluation.

Results—Eighteen studies meeting inclusion criteria were identified. The reported frequency of malnutrition ranged from 6.1% to 62%. Seventeen different methods of nutritional assessment were used. Four trials used previously validated assessment methods: Subjective Global Assessment, "an informal assessment," and Mini Nutritional Assessment. The nutritional assessment methods used in the remaining studies used had not been validated previously.

Conclusions—The use of a wide assortment of nutritional assessment tools, many of which have not been validated, may have contributed to the wide range of estimates of malnutrition. If so, this underscores the need for valid and reliable assessment tools to further our understanding of the relationship between stroke and nutritional status.