If the aim of nutritional assessment of preterm infants is to identify suboptimal (or excessive) provision of protein, energy and micronutrients, most currently available methods perform poorly. Assessment of body weight is limited by the confounding effect of fluid status especially in the first few days of life, and measurements of linear growth are relatively imprecise and slow to respond to nutritional changes. Growth assessment is hampered by the lack of an adequate reference standard. Comparisons to historical cohorts of preterm babies are inadequate. As most very low birth weight infants leave hospital below the 10th centile, use of these charts as 'standards' almost guarantees that preterm infants will have poor growth. Growth centiles based on data from newborn preterm infants have certain advantages. However, this is hardly normative data as preterm birth is always an abnormal event. Methods of assessing body composition are largely limited to the research setting, and it remains unclear whether the optimum composition of postnatal growth is one that mimics fetal growth or postnatal growth of the term infant. Biochemical nutritional assessments are of limited utility except in the highest-risk preterm infants, when nutritional inadequacy is likely (severe fluid restriction) or where intake is difficult to assess (use of human milk).