Results
Two hundred and seventy-four women were invited, and 264 (95%) women agreed to participate. Of the ten women (5%) who declined to patriciate: six women did not provide any reason, one reported in pain, one was not interested and two declined because their partner did not want the women to participate. Two hundred and twenty-six women delivered vaginally. Two hundred and nineteen were nulliparous, six (2.7%) had had a previous caesarean section and one (0.5%) had had two previous caesarean sections (none of which were in the second stage of labour). One hundred and twenty-one (53.5%) perineal tears were examined by midwives and 105 (46.5%) by doctors (Table 1). Twelve (4.6%) women had an intact perineum, eight (3%) a first-degree tear and 185 (70.3%) a second degree tear. Twenty-one (9.3%) women sustained OASIs: eight (3%) a 3a tear and 13 (4.9%) a 3b tear. Of these, five women were examined by midwives who missed four (80%) (Table 1) and sixteen were examined by doctors who missed two (12.5%) OASIs (Table 2).
Two hundred and eighteen of the 226 (96%) women had ultrasound images that were of a suitable quality for analysis. In 21 (9.6%) women an anal sphincter defect was identified by TPUS. Table 3 shows the correlation between ultrasound and clinical findings. Of the 21 OASIs that occurred 19 (90.5%) women had an anal sphincter defect identified with TPUS at delivery. No woman had an IAS injury that was diagnosed clinically or with TPUS (Table 3). In addition, there were two (1%) defects seen on ultrasound that were not clinically detectable. The diagnostic performance of TPUS for detecting OASIs was evaluated, the positive and negative predictive of TPUS being 91% and 99% respectively (Table 4).
The relationship between the size of EAS defects identified clinically and by ultrasound was also analysed. There was no correlation between the grade of OASIs on clinical examination and on TPUS (χ2 test, p=0.14) (Table 5).