OBJECTIVE In an earlier study we have shown that transcervical chorionic villus sampling in excess of 90 mg increases the risk for hemangiomas of infancy three- to four-fold compared to amniocentesis. In the present study we investigated whether transabdominal chorionic villus sampling (TA-CVS), in which the samples are smaller, carries the same risk. METHODS Retrospectively, data were analyzed from 200 consecutive TA-CVS procedures and 200 consecutive amniocentesis procedures. Forty-two TA-CVS procedures and 27 amniocentesis procedures were excluded on predefined criteria. Questionnaires were sent to the parents asking if there was any skin mark on the child: vascular, pigmented or otherwise. All hemangiomas were clinically confirmed. RESULTS In the TA-CVS group, 118/158 questionnaires (75%), and in the amniocentesis group 134/173 questionnaires (77%) were returned. Based on the results of the questionnaire (i.e. mentioning of any skin lesion), 24 children in the TA-CVS group and 42 children in the amniocentesis group qualified for a physical examination. In the TA-CVS group 11/118 children (9%) had one or more hemangiomas. In the amniocentesis group 6/134 children (4%) had one or more hemangiomas. There was no statistical difference between the two groups (P = 0134). CONCLUSIONS These results suggest that TA-CVS does not cause an increase in the prevalence of hemangioma compared to amniocentesis. A larger series is, however, necessary to confirm this.