Objective
We sought to determine factors associated with perineal wound complications in women with obstetric anal sphincter injuries (OASIS).
Study Design
A retrospective chart review of women who sustained an OASIS from Nov. 2, 2005, through March 1, 2010, was performed.
Results
In all, 1629 women sustained an OASIS; 909 had follow-up data. Wound complications (infection, breakdown, packing, operative intervention, secondary repair) occurred in 7.3% (n = 66) of patients. Smoking (odds ratio [OR], 4.04; 95% confidence interval [CI], 1.4–12.2; P = .01), increasing body mass index (OR, 1.06; 95% CI, 1.01–1.12; P = .04), fourth-degree laceration (OR, 1.89; 95% CI, 0.99–3.61; P = .05), operative vaginal delivery (OR, 1.76; 95% CI, 1.15–2.68; P = .009), and use of postpartum antibiotics (OR, 2.46; 95% CI, 1.11–5.63; P = .03) were associated with complications; intrapartum antibiotics were protective (OR, 0.29; 95% CI, 0.14–0.59; P = .001). In all, 44% of patients with a complication (n = 29) required hospital readmission; most (72%) were in the first 2 postpartum weeks.
Conclusion
Wound complications after OASIS are associated with modifiable factors and often require hospital readmission.
Complications following obstetric anal sphincter injuries (OASIS) during vaginal delivery are relatively rare events, with an incidence of approximately 5-13%. One of the most devastating complications is perineal wound breakdown, occurring with an incidence of 0.1-4.6%. Although uncommon, perineal wound complications can lead to significant morbidity, including chronic pain, incontinence, embarrassment, rectovaginal fistula, and loss of sexual function. Many risk factors for third- and fourth-degree perineal lacerations have been identified, including race, operative vaginal delivery, episiotomy, perineal length, nulliparity, length of second stage, and birthweight. However, there has been relatively little research to identify factors associated with these perineal wound complications.
At our tertiary care institution in 2009, 382 women sustained an OASIS, for an overall rate of 4.5%. The rate reported in the literature has been quite variable, ranging from 0.25–6%. As higher-order perineal lacerations are associated with an increased risk of morbidity, it is essential to identify factors associated with complications in this group of women, so that we might be able to implement preventive techniques to avoid such devastating outcomes.
The purpose of the current study was to identify demographic, antepartum, intrapartum, and postpartum factors associated with perineal wound complications. By obtaining these data, we hope to be able to identify modifiable labor practices, better counsel pregnant women, and refine our postpartum care to decrease the rates of both the short- and long-term sequelae of severe lacerations.
Materials and Methods
This was a retrospective chart review of all women who had a vaginal delivery complicated by an OASIS at Prentice Women’s Hospital, Chicago, IL, from Nov. 2, 2005, through March 1, 2010. Patients were identified using the Northwestern University Medical Enterprise Data Warehouse (EDW), an electronic repository of all data from our electronic medical records systems. The degree of perineal laceration sustained is coded within this EDW. Additional data not hard-coded in the EDW including length of ruptured membranes, length of second stage of labor, and suture material utilized for repair of the laceration were collected via chart review by the authors.
The sole inclusion criterion was having sustained an OASIS during the aforementioned time period. Twin deliveries were excluded from analysis. We did not exclude preterm deliveries, or patients with preterm premature rupture of membranes or other antepartum issues. Third- and fourth-degree lacerations were identified by the delivering attending physician and resident physician. Tears coded as a “partial” third-degree were included.
Antepartum, intrapartum, and postpartum variables were assessed using the EDW.
The study was approved by the Northwestern University Institutional Review Board. SPSS (version 18.0; SPSS Inc, Chicago, IL) was used for all data analyses. Categorical variables were evaluated using a χ 2 analysis and continuous variables were evaluated using 2-tailed, paired t tests. Binomial logistic regression models were used to study the extent to which individual variables or combinations of variables predicted wound complications. Alpha was set to 0.05 for all analyses.
Results
In all, 1657 patients were identified by the EDW as having sustained an OASIS. A total of 28 patients were excluded due to twin gestations, for an ultimate sample size of 1629. Of this cohort, postpartum follow-up data were available for 909 subjects ( Figure 1 ). These 909 patients were women who received all of their care from practices using the electronic medical records for both inpatient and outpatient care. The 720 patients for whom complete data were not available received their antenatal and postnatal care from multiple, private, community practices using paper charting; thus, while their inpatient hospitalization could be captured, it was not possible to reliably ascertain postpartum complications, which might have been handled in the outpatient setting.
The demographics of the cohort are listed in Table 1 . The majority of the cohort was Caucasian (71.3%), nulliparous (92.8%), and had a mean age of 31.8 (±4.3) years. About 20% of women were colonized with group B streptococcus (GBS).
| Demographic variables | n (%) |
|---|---|
| Age, y, mean (SD) | 31.8 (4.3) |
| Race | |
| Caucasian | 648 (71.3) |
| African American | 36 (4.0) |
| Asian | 45 (5.0) |
| Hispanic | 62 (6.8) |
| Pacific Islander/Hawaiian | 34 (3.7) |
| Other/missing | 84 (9.2) |
| Nulliparous | 844 (92.8) |
| Smoker | 19 (2.1) |
| BMI, kg/m 2 , mean (SD) | 28.8 (4.4) |
| Gestational DM | 17 (1.9) |
| Preexisting DM | 4 (0.4) |
| Group B streptococcal colonization | 186 (20.5) |
Pertinent intrapartum data are shown in Table 2 . About 16% of women were diagnosed with chorioamnionitis during their labor course, and 34% of the cohort was given antibiotics in labor secondary to chorioamnionitis, GBS carrier status, or both. Almost two-thirds of the women in this cohort underwent an assisted vaginal delivery; the most common indication for forceps and vacuum use was maternal exhaustion. Very few women experienced a postpartum hemorrhage. In 78% of subjects (n = 711) ( Table 2 ), providers specifically documented type of suture used and most lacerations were repaired using only polyglactin suture.
| Delivery variables | Mean (SD) or n (%) |
|---|---|
| Birthweight, g | 3534.3 (448.6) |
| Chorioamnionitis | 143 (15.7%) |
| Antibiotics in labor | 308 (33.9%) |
| Meconium-stained fluid | 269 (29.6%) |
| Length of ruptured membranes, min | 685.7 (610.8) |
| Length of second stage, min | 137.5 (88.8) |
| Shoulder dystocia | 59 (6.5%) |
| Route of delivery | |
| Spontaneous | 336 (37.0%) |
| Forceps assisted | 496 (54.6%) |
| Vacuum assisted | 77 (8.5%) |
| Episiotomy | |
| Midline | 175 (19.3%) |
| Mediolateral | 19 (2.1%) |
| Degree of laceration | |
| Third degree | 784 (86.2%) |
| Fourth degree | 125 (13.8%) |
| Indication for operative delivery | |
| Exhaustion | 257 (44.7%) |
| Arrest of descent | 189 (32.9%) |
| Nonreassuring fetal tracing | 181 (31.5%) |
| Other/not documented | 15 (2.5%) |
| Estimated blood loss >500 mL | 100 (11%) |
| Suture material utilized (n = 711) | |
| Polyglactin only | 616 (86.6%) |
| Chromic only | 11 (1.5%) |
| Polyglactin and chromic | 84 (11.8%) |
Wound complications occurred in 66 (7.3%) of the patients in the cohort. In all, 39 women (59.1%) had a documented wound infection, 36 (54.5%) had a breakdown of their wound, 17 (25.8%) required reopening of the wound, and 8 (12.1%) required packing of the wound. Over 50% of wounds (34/66) required operative intervention including debridement; 24 (36.4%) of these returned to the operating room for a secondary repair. Patients often had >1 of these complications concurrently.
The factors associated with wound complications in univariate analysis are shown in Table 3 . Demographic variables of race, body mass index (BMI), and smoking were significantly associated with wound complications, while medical comorbidities were not. Women who were colonized with GBS were more likely to have a wound complication as compared to those who were not colonized. Women with a fourth-degree laceration (compared to third-degree), those who sustained a blood loss of >500 mL, and those receiving antibiotics in the postpartum period were also more likely to have wound complications. The use of intrapartum antibiotics was found to be associated with a reduced risk of wound complications.
| Variable | No complications, n (%) | Complications, n (%) | P value a |
|---|---|---|---|
| Demographic variables | |||
| Race | .001 | ||
| Caucasian | 605 (93.6) | 42 (6.4) | |
| African American | 34 (97.1) | 1 (2.9) | |
| Asian | 42 (93.3) | 3 (6.7) | |
| Hispanic | 53 (85.6) | 9 (14.5) | |
| Pacific Islander | 26 (76.5) | 8 (23.5) | |
| Decline | 81 (96.4) | 3 (3.6) | |
| Age, y, mean (SD) | 31.6 (4.5) | 31.9 (4.2) | .55 |
| BMI, kg/m 2 , mean (SD) | 30.0 (5.1) | 28.7 (4.3) | .03 |
| Smoking | 5 (27.8) | 13 (72.2) | .001 |
| Nulliparity | 61 (7.2) | 783 (92.9) | .89 |
| Medical comorbidities | |||
| Type 2 diabetes | 0 (0) | 4 (100) | .58 |
| Gestational diabetes | 3 (17.6) | 14 (82.4) | .1 |
| Any diabetes | 3 (14.3) | 18 (85.7) | .21 |
| Asthma | 4 (6.4) | 58 (93.6) | .92 |
| Inflammatory bowel disease | 0 (0) | 5 (100) | .55 |
| Intrapartum variables | |||
| Meconium-stained fluid | 22 (7.3) | 247 (92.7) | .49 |
| GBS colonization | 6 (3.2) | 180 (96.8) | .04 |
| Chorioamnionitis | 8 (5.6) | 144 (94.4) | .4 |
| Antibiotics during labor | 11 (3.5) | 299 (96.5) | .005 |
| Shoulder dystocia | 4 (6.8) | 55 (93.2) | .88 |
| Birthweight, g | 3566 (504) | 3531 (444) | .55 |
| Estimated blood loss >500 mL | 12 (12) | 88 (88) | .05 |
| Episiotomy | 9 (4.6) | 185 (95.4) | .2 |
| Type of vaginal delivery | 13 (3.9) | 323 (96.1) | .009 |
| Spontaneous | 47 (9.5) | 449 (90.5) | |
| Forceps | 6 (7.8) | 71 (92.2) | |
| Vacuum | |||
| Laceration variables | |||
| Fourth-degree laceration | 16 (13.0) | 107 (87.0) | .003 |
| Postpartum antibiotics | 10 (6.3) | 564 (91.6) | .05 |
| Suture type utilized | |||
| Polyglactin | 52 (8.4) | 11 (100) | .31 |
| Chromic | 0 (0) | 80 (95.2) | |
| Both polyglactin and chromic | 4 (4.8) | 69 (92.7) |
a Categorical variables were analyzed using χ 2 analysis; continuous variables were analyzed using t tests.
Method of vaginal delivery was a significant predictor of wound complications. More specifically, wound complications occurred in 9.5% vs 7.8% vs 3.9% of forceps, vacuum, and spontaneous deliveries, respectively ( P = .009). To account for operative vaginal delivery potentially serving as a surrogate for a prolonged second stage of labor, method of delivery was reanalyzed while controlling for length of the second stage; method of vaginal delivery remained a significant predictor of wound breakdown ( P = .023). Furthermore, length of second stage was not associated with rate of complications after OASIS in univariate analysis.
A model to predict perineal wound complications was then constructed using multivariable binary logistic regression. Variables were added to the model in a step-wise fashion and were included in the most parsimonious final model if they significantly improved its predictive value. A test of the full model vs a model with intercept only was statistically significant: χ (6, N = 892) = 37.7, P < .0001. The model was able to correctly classify women 92.9% of the time ( Table 4 ). Employing a .05 criterion of statistical significance, BMI, smoking, operative vaginal delivery, degree of tear, and antibiotics each had significant partial effects. Intrapartum antibiotics were protective against wound complications, whereas postpartum antibiotics were associated with an increased risk of complications.
