We read with interest de Leeuw et al.’s large retrospective study evaluating the relationship between episiotomy and anal sphincter tears during forceps and vacuum deliveries.1 The results were surprising and very different from other previously published studies. First, the rate of anal sphincter injuries raises concerns about underascertainment or underreporting. The majority of deliveries were performed with the use of mediolateral episiotomy, which is similar to the practice in the UK, particularly for forceps delivery. de Leeuw et al.’s study reported an anal sphincter tear rate of 1.4% for vacuum deliveries and 2.6% for forceps deliveries. These are substantially lower than most reported rates and conflict with the well-recognised association between operative vaginal delivery and an increased risk of anal sphincter tears.2,3 Second, the magnitude of association between operative vaginal delivery without episiotomy and anal sphincter injury is extremely high. We question why obstetricians in the Netherlands continue to perform forceps delivery without episiotomy if anal sphincter tearing occurs once in every five cases (especially in the context of a very low background rate). Our findings in a Scottish study were very different for both vacuum and forceps.4 The database we used contained records completed by both the midwife and the obstetrician. In some parts of the database, only the most serious morbidity was recorded, i.e. third- or fourth-degree tear rather than all morbidities, including use of an episiotomy. We cross-referenced all recorded variables on perineal trauma to identify inconsistencies and where necessary reverted to the handwritten records. This gave us an accurate database of the procedures used and the outcomes. The findings of the Dutch study suggest that there has been incomplete recording of all types of perineal morbidity at operative vaginal delivery, with a default to record only the most serious morbidity. We would welcome clarification on these issues and would urge caution with regards to the authors’ forceful conclusion, given the known limitations of retrospective studies.