BACKGROUND: Amniocentesis in the third trimester, which reduces risks of procedure-related miscarriage but still allows termination of affected fetuses, may be applicable in some pregnancies. The implications of deferring amniocentesis include complications, delivery before the test and increased amniotic fluid culture failure rates. We investigated the indications, complications, karyotype results and laboratory failure rates of third-trimester amniocentesis. METHODS: We studied all women who underwent third-trimester amniocentesis from 2000 to 2006. Data were collected from ultrasound databases, computerised records and individual chart review. RESULTS: We reviewed 165 pregnancies that underwent amniocenteses after 28 weeks. Median maternal age at amniocentesis was 32 years and median gestation, 32(+2) weeks. Indications included malformation (60/165), soft markers (37/165), maternal request (12/165), and positive screening test (11/165). Of the 49 women(29.7%) who declined second-trimester amniocentesis, 24.5% had twins and 38.8%, malformations. Amniocentesis was not offered to 116 women: 57/116 (49.1%) third-trimester referrals, 25/116 (21.5%) diagnosed late and the remainder, low-risk indications. Fetal karyotype was abnormal in 17 cases (10.3%). Seven women who initially declined amniocentesis had abnormal results compared with one advised to have late amniocentesis. Culture failure rate was 9.7%, however results were obtained by Quantitative fluorescent polymerase chain reaction (QF-PCR) from 164/165 samples. Complication rate was 1.2%. CONCLUSION: For late diagnoses and for low-risk indications, third-trimester amniocentesis is an acceptable option, especially when utilising QF-PCR with cytogenetic culture.