PURPOSE:: About 50% of spontaneous abortions are caused by fetal chromosome abnormalities. Identification of these abnormalities helps to estimate recurrence risks in future pregnancies. However, due to culture failures or maternal contamination often no fetal karyotype can be obtained. Array comparative genomic hybridization can overcome some of these limitations. METHODS:: In this study, we analyzed 103 miscarriages by both T-banding and 1-Mb array comparative genomic hybridization. RESULTS:: We found an overall abnormality rate of 35% (34 of 96). In a comparison of 70 samples that were successfully analyzed by both techniques, 54 (77%) had identical karyotypes (42 normal, 12 abnormal) and 16 (23%) cases showed discrepancies. Most of these differences were due to maternal contamination during cell culture, which resulted erroneously in a normal female karyotype. CONCLUSION:: These results demonstrate the improved diagnostic yield of array comparative genomic hybridization as compared with conventional karyotyping. Therefore, we implemented this technique in the diagnostic workup of miscarriages.