Hypothesis / aims of study
Our objectives were threefold: To (1) identify risk factors associated with obstetric anal sphincter injuries (OASIS) in women undergoing vaginal birth after cesarean section (VBAC), (2) determine if a relationship exists between predicted probability of VBAC success and OASIS and (3) develop a prediction model for OASIS in women who are undergoing VBAC.
Study design, materials and methods
This was a retrospective case-control study at a single center. Participants had a singleton VBAC between January 2011 and December 2016. Cases were women who sustained OASIS at the time of VBAC. From electronic medical records, we extracted subjects’ demographic data, obstetric, and medical history (maternal age, height, body mass index (BMI), reported race or ethnicity, parity, smoking status, indication for prior cesarean section, history of any prior vaginal delivery, history of hypertension and/or diabetes). We also collected data on subjects' intrapartum course and delivery characteristics. Predicted probability of VBAC success was calculated using the Maternal Fetal Medicine Units (MFMU) Network Vaginal Birth After Cesarean calculator [1]. Univariate analyses were performed using statistical analyses as appropriate to identify antepartum and intrapartum variables significantly associated with OASIS. These variables were then used to create the most parsimonious prediction model using a significance threshold of 0.15 for inclusion and backward stepwise logistic regression. A second model was generated by substituting the VBAC success in place of variables used in its generation. Pseudo R2 and AIC values were used to compare models. Data was analyzed using Stata 11.2 (Stata- Corp, TX, USA) and SPSS (Version 20, Chicago IL).
Results
1411 women met inclusion criteria. 73 (5.2%) sustained OASIS at the time of VBAC; 2.8% occurred with spontaneous VBAC, 30.1% with forceps-assisted VBAC and 13.3% with vacuum assisted VBAC, p=0.001 (Table 1). On univariate analysis, OASIS was associated with operative vaginal delivery and episiotomy (OR 12.74, 95% CI 7.72 - 21.03 p<0.001 and OR 3.79, 95%CI 1.71-8.41, p= 0.003 respectively), while African American race and predicted VBAC success probability of >75% were associated with decreased odds of OASIS (OR 0.22, 95% CI 0.71- 0.73 and OR 0.43, 95% CI 0.25- 0.76, p<0.05 respectively). On multivariable logistic regression, African-American race, episiotomy and operative vaginal delivery maintained significance(Table 2). Substitution of the predicted VBAC success did not improve model goodness of fit. The model demonstrated that African American race (OR 0.17, CI 0.05- 0.58) was protective against OASIS. In contrast, episiotomy (OR 3.39, 95% CI 1.34- 8.58), forceps delivery (OR 15.73, 95% CI 9.14-27.05) and vacuum delivery (OR 3.59, 95% CI 1.07-12.0) were significant risk factors for OASIS (Table 2). Our model also demonstrated that the use of forceps increased the probability of OASIS by 10-fold in Caucasian women who undergo VBAC (3.4% vs 35%).
Interpretation of results
In women who undergo VBAC the use of forceps conferred the highest odds for OASIS. Forceps delivery at the time of VBAC carries up to a 30% increase risk of OASIS. African American race and predicted probability of VBAC success greater than 75% were protective against OASIS at the time of VBAC.