In our study, the incidence of perineal tears was not modified in case of a spontaneous delivery following a fetal demise. No OASIS were found in the two groups. Stillbirth was neither a protective factor nor a risk factor for any perineal tears. Only a trend of a higher rate of second-degree perineal tear was highlighted.
These results are not consistent with the Basu et al.9 study who compared the incidence of perineal tears between 323 women who had a spontaneous vaginal delivery following a fetal demise and 1000 women who delivered a live-born child. In their study, IUFD reduced the relative risks of any perineal tears (RR = 0.16 95% CI [0.12–0.22]) and OASIS (RR = 0.12 95% CI [0.03–0.55]). One explanation for this difference could be the very low incidence of OASIS in our study population that is about 0.5% per year for all spontaneous and instrumental vaginal deliveries. During the study period, these OASIS mainly occurred during instrumental deliveries. The rate of instrumental delivery in fetal demise deliveries is about 11% in our center. Our team uses mainly the vacuum in live birth and spatulas or forceps in fetal still birth. The two groups would not have been comparable. In the same way, no prolonged second stage of labor and prolonged expulsive efforts were found in our data collection. These elements are known to be risk factors for OASIS. In addition, our study was conducted in a level III maternity hospital with an important policy of perineal protection with the realization of a systematic Couder’s maneuver (78% of vaginal deliveries) and a selective use of episiotomy (0.01% per year).
In our study, there was a trend toward a lower incidence of second-degree perineal tear in the IUFD group (18% versus 28%, RR = 0.51 95% CI [0.2–1.4]). This could be explained by the fact that the macerated demise fetal head is more easily deformed during delivery. Therefore, there are less stress on the perineum which would induce less perineal tears. But it has never been studied biomechanically in the literature.
The two groups were not comparable in terms of analgesia (76% versus 98% in the IUFD group, p < 0.01). Indeed, women who underwent vaginal deliveries following a fetal demise had more epidural analgesia according to the department protocol and in order to reduce the women pain. However, this element was not a confounding bias in our study. Indeed, it has been shown by Loewenberg et al. 10 that epidural analgesia was not a risk factor for severe perineal tear.
Regarding the occurrence of first- and second-degree lacerations and analgesia in the literature, no statistically significant association seemed to exist. Bodner-Alder et al. 11 showed no evidence of a detrimental effect of the epidural analgesia on the integrity of the birth-canal.
The rate of induction of labor was not comparable between the two groups (86% vs 17%, p < 0.01). Women in the IUFD group who were not in spontaneous labor had an induction. According to Grobman et al.12 the incidence of perineal tears did not differ between the group of women who delivered after a labor induction and those in which the labor was spontaneous. The duration of expulsive efforts was significantly longer in the control group (5 min versus 7 min, p < 0.01). But this little difference of two minutes was not clinically significant because it remained short. The second stage of labor was statistically longer in the control group (69 min versus 31 min, p < 0.01). This difference could be explained by an absence of fetal tone and a trend of smaller fetal head diameter due to maceration. These factors are involved in the biomechanics of labor and delivery. Indeed, Lipschuetz et al.13 demonstrated that a high head circumference increased the duration of the second stage of labor. Valsky et al.14 also demonstrated that a high fetal head circumference was a risk factor for perineal tears. Known as a risk factor for severe injuries, the shorter duration of the second stage of labor in the IUFD group may explain the lack of OASIS and the trend toward a lower incidence of second-degree perineal tears.
There was no statically significant difference in head circumference measurements between the two groups (28 cm versus 30 cm in the control group, p = 0.08). But there was a trend of lower values in the IUFD group with a difference close to significance (p = 0.08). This trend was also found by Pacora et al.15 who found significantly lower cranial perimeters on pre-mortem ultrasound in the “fetal death” group compared to the “control” group with live fetus. This may be explained by some causes of in utero fetal death (intrauterine growth retardation or genetic abnormalities). This phenomenon is also associated with a decrease in these diameters in postmortem by maceration of the tissues, reducing also the trophicities of the latter. Furthermore, a relationship between the tissue stretches and the fetal head diameter was shown by Lien et al. A difference of head circumference close to significance could lead to an increase of perineal stretch, and so, to perineal injuries16.
Silva et al.17 studied the influence of the fetal head molding on the biomechanical behavior of the pelvic floor muscles. During a vaginal delivery, the forces applied to the fetal head by the pelvic floor induced a plastic deformation of the head and a 17.3% reduction in the reaction forces on the pelvic muscle floor. Fetal heads of deceased fetus would have a greater capacity for deformation due to the absence of tonicity and a smaller head circumference. Therefore, this would reduce the reaction forces of the head on the muscle floor and thus induce a lower incidence of second-degree perineal injuries18.
Lien et al., studied the stretch of the ani levator muscle during vaginal delivery. They demonstrated that the relationship between the tissue stretch, and the fetal head diameter was proportional. A reduction in this diameter would therefore lead to a decrease in the stretch of the levator ani muscle16. The same rationale could be applied to the perineal stretch.
Our study concerned a large cohort of women who delivered over fourteen successive years. One of the strengths of this work was the matching on the main confounding factors of perineal tears. There was no matching on induction because the methods of induction of labor are not identical in cases of in utero fetal death. The experience of the accoucheur was similar between the two groups because all women were delivered by the same experienced midwives. The information bias related to the retrospective nature of this study was minimized by using a database that was exhaustively completed by the medical team in the maternity ward. There were no missing data. In contrast to the study by Basu et al.9, maternal overweight was considered in this work to remove this confounding bias. There could be a bias in the classification of perineal tears due to the retrospective nature of this work. However, we consider this error rate to be very low due to double checking (midwife and senior) in case of doubt in the classification between a second- or third-degree perineal tear.
This lack of difference of perineal injuries during a fetal demise or a live birth delivery can be explained by the same policy of perineal protection that our center applies to every delivery. This policy can also explain the difference between our study and Basu et al., study. Indeed, the overall rate of perineal trauma was 59.5% in the live birth group from Basu et al.9 study versus 16% in our control group. Our perineal protection policy seems to be more efficient19,20,21.
But biomechanical studies are necessary to better understand perineal stretching and the risk of perineal rupture according to mechanical constraints related to the maceration of the fetal tissues.