Fetal death in a twin conception during second and third trimester is associated with increased risk of cerebral injury in the surviving twin. The aim of this study is to test the hypothesis that even early fetal losses as a ‘vanishing’ twin may be associated with an increased risk of cerebral impairment in the surviving twin. The study population comprised 362 pregnant women attending Liverpool Women's Hospital recruited between 1999 and 2001. Women were classified according to the first ultrasound scan into 3 groups: vanishing twin, twin and singleton. The vanishing twin group was further subdivided into ‘definite’ and ‘probable’. Children from these pregnancies were assessed at 1 year of age for their development and neurological function using the Griffiths Mental and Developmental Scales and Optimality score. Children from 229 pregnancies (63.2%) attended the assessment. Information on children from a further 21 (5.8%) pregnancies was obtained through a review of hospital records. Cerebral impairment was found in 2 children from the vanishing twin group, 2 from the twin group and none from the singleton group. When cases with definite vanishing twin are considered there is a significant difference between the vanishing twin and singleton group (relative risk 6.1; 95% confidence interval 1.5–8.3; p = .03). An additional study with an increased sample size would enable a more robust conclusion.