Effect of the “shoulder-up” bundle on the incidence of spontaneous perineal injury after vaginal delivery: comparison of 2 historic cohorts after propensity score matching

Perineal injury following vaginal delivery represents a major cause of long-term maternal morbidity, and its prevention is among the priorities of modern obstetrical practice. This study aimed to investigate whether the systematic implementation of a bundle of maneuvers to prevent perineal injury (ie, the “shoulder-up” bundle) may reduce the rate of spontaneous perineal tears in women delivering at a single tertiary maternity unit. This was a single-center retrospective intervention study including all vaginal deliveries between April 1, 2020 and March 31, 2022. On March 1, 2021, a bundle focused on perineal injury prevention in vaginal deliveries was implemented and introduced as a standard of care. The “shoulder-up” bundle includes the addition of a hands-on technique for the delivery of the posterior shoulder, which is slowly lifted up under the direct visualization of the perineal body, immediately after the disengagement of the anterior shoulder. The labor ward staff underwent dedicated training to acquire expertise on the “shoulder-up” bundle. Little changes in terms of medical and midwifery staffing were recorded during the study period. The incidence of spontaneous second-degree or higher perineal tears was compared between the patients who gave birth before the clinical implementation of the bundle (standard-care group) and those who were delivered following the implementation of the bundle (shoulder-up group). A 1:1 propensity score matching of the 2 groups was done for the variables that proved to be independently associated with the perineal outcome. From April 1, 2020 to March 31, 2022, 3671 patients had a vaginal birth at our tertiary care unit (1786 in the standard-care group and 1885 in the “shoulder-up” group) and were enrolled in the study population. Of these, 1191 (32.4%) had a spontaneous second-degree or higher perineal tear. At univariate analysis, nulliparity (59.6% vs 39.1%; P