Objective-To assess the relation between the indication for fetal blood sampling and pregnancy loss following the procedure. Design-Retrospective study. Setting-The tertiary referral Fetal Mcdicine Units at Guy's and University College Hospitals, London. Si4bjects-Women undergoing diagnostic fetal blood sampling in four groups:(1) 94 having prenatal diagnosis with normal ultrasound findings; (2) 94 with a structural fet31 abnormality; (3) 30 having fetal assessment and (4) 35 with nonimmune hydrops. Inferventiotzs-Freehand ultrasound guided fetal blood sampling from umbilical cord, intrahepatic vein or fctal heart. Main oritcome measiires-Pregnancy losses were divided into those within 2 weeks and thosc 2 wccks aftcr the procedure, obstetric accidents and nconatal deaths. Resiilts-The 253 patients had fetal blood samplcd o n 268 occasions. Fifty-one pregnancies were terminated. Overall, 5 1 of the remaining 202 desired continuing pregnancies were lost, of which 19 (9%) were lost within 2 weeks of the procedure. After exclusion of the pregnancies that were terminated, the procedure-related losses within 2 wecks of sampling wcre 1 in 76 (1 %), 5 in 76 (7%). 4 in 29 (14%) and 9 in 36 (25%) in groups I , 2, 3 and 4 respectively. Cotdusions-The risk of fetal blood sampling is incrcascd in abnormal pregnancies, reflecting the underlying pathology and this must be taken into account when counselling patients before thc procedure.

publication will not be returned. Authors whose papers are rejected will be advised of the decision, and the manuscripts will be kept under security for three months, to deal with any inquiries, and then destroyed by shredding. Hence we would prefer to receive for consideration photostats or copies produced by word processor (see BMJ 13 October, p 942), though we do, of course, still need three copies.

This report describes the early prenatal diagnosis of the Pena Shokeir phenotype in an at-risk patient at 14 weeks' gestation. The diagnosis was based on an abnormal fetal movement profile, in association with an abnormal position of the fetal limbs. Pena Shokeir phenotype describes an inherited condition characterized by arthrogryposis and dysmorphic features as a result of fetal akinesia. It is a lethal abnormality and early diagnosis allows safer surgical methods of termination.

The red cell alloantibody, anti-U, is uncommon but is a recognised cause of haemolytic disease of the fetus and newborn. We describe six pregnancies complicated by the presence of maternal anti-U, and review nine other well-documented cases. In these 15 cases severe haemolytic disease occurred only with titres of 2 1/512, and titres as high as 1/4000 were not necessarily associated with significant haemolysis. We recommend that an anti-U titre of 2 1/128 or more at 2 17 weeks of gestation is an indication for assessment of haemolysis in the fetus. Amniocentesis is the preferred initial investigation.

Intracerebral tumours of the fetus are very rare conditions, most often presenting clinically as polyhydramnios and hydrocephalus. These conditions can be diagnosed with ultrasound and clearly differentiated from hydrocephalus and other intracranial lesions. The following report is of a case of an intracerebral tumour (glioblastoma multiforme) diagnosed at 33 weeks in utero using ultrasound. The prognosis for this condition is universally poor.