Obstetric Myths Versus Research Realities
Chapter 14: Episiotomy
Episiotomies Do Not Perform as Advertised (Reviews)
Episiotomies Do Not Perform as Advertised (Studies)
Episiotomies Do Not Prevent Deep Tears
Episiotomies Do Not Heal Better Than Tears
Episiotomies Do Not Prevent Pelvic Floor Muscle Relaxation
Episiotomies Do Not Prevent Fetal Brain Damage
Episiotomies Do Not Perform as Advertised (Reviews)
1. Thacker SB and Banta HD. Benefits and risks of episiotomy: an interpretive review of the English language literature, 1860-1980. Obstet Gynecol Surv 1983;38(6):322-338.
The authors reviewed more than 350 articles, reports, and book chapters published between 1860 and 1980 and found no convincing evidence that episiotomy prevented tears into the rectum, damage to the pelvic wall, or trauma to the fetal head or that episiotomies were easier to repair than tears. Like any other surgery, episiotomy has risks, including extension of the incision, blood loss, dyspareunia, pain, poor healing, and infection, including fatal infection, all of which are documented in the literature. "[T]he risks of episiotomy are more severe than many might appreciate. Although rarely associated with a life-threatening problem, the complications of this procedure can be a source of serious morbidity to young mothers who already have major personal and social adjustments to undergo." In some cases, episiotomy is justified but not as a routine procedure. Based on rates reported at birth centers and home births, 20% seems a reasonable episiotomy rate. "If [women] were fully informed as to the evidence for benefit and risk in the face of demonstrable risks, it is unlikely that women would readily consent to having routine episiotomies."
2. Bromberg MH. Presumptive maternal benefits of routine episiotomy. J Nurse Midwifery 1986;31(3):121-127.
"A survey of the available literature is remarkable for the astonishing dearth of valid data to support long-held beliefs about episiotomy." This review focuses on two of them: preservation of pelvic floor integrity and prevention of lacerations.
Episiotomy is supposed to prevent overstretching of the muscles, which is believed to be the cause of poor perineal tone. (Poor perineal tone increases the risk of cystocele, rectocele, uterine prolapse, and stress incontinence.) What little evidence exists in favor of episiotomy is over 60 years old and of poor methodology. Some causes of pelvic floor relaxation may be iatrogenic: poor episiotomy repair or pulling the cervix down for inspection [see DeLee's protocol above]. As for prevention of lacerations, at best women do no worse without an episiotomy with respect to pain and lacerations, and some studies show an association between episiotomy and deep tears. Iatrogenic factors contributing to lacerations may include: use of techniques to shorten second stage (e.g. fundal pressure or strenuous pushing), failure to wait for shoulder rotation before delivering them, faulty repair of a previous episiotomy, and possibly injecting local anesthetic but not doing an episiotomy. "Review of the literature on episiotomy indicates the likelihood that it is over used, with shaky justification at best. It seems reasonable to infer that a median episiotomy has no great advantage over a first- [into the skin] or second-degree [into the underlying muscle] laceration when there are no overriding fetal indications." [An episiotomy is equivalent to a second-degree laceration.]
3. Hofmeyr GJ and Sonnedecker EW. Elective episiotomy in perspective. S Afr Med J 1987;71(6):357-359.
This article summarized recent clinical studies and concluded that episiotomy did not prevent genital prolapse or fetal damage, was not easier to repair than a second-degree tear, did not heal better, and did not prevent third-degree laceration or impairment of coital function. The authors observed that if episiotomy is to prevent overstretching of the perineum and protect the fetal head, it must be done early, but episiotomies are not done until maximal stretching has already occurred. "If elective episiotomy does indeed confer any of the benefits conventionally ascribed to the procedure, properly controlled evidence for such benefit is yet to emerge. . . . It is the responsibility of the proponents of elective episiotomy to provide such evidence."
4. Thorp JM and Bowes WA. Episiotomy: Can its routine use be defended? Am J Obstet Gynecol 1989;160(5 Pt 1):1027-1030.
The literature is reviewed for support of the two main rationales for routine episiotomy: reduction of perineal trauma and prevention of pelvic musculature relaxation. No study found that midline or mediolateral episiotomy reduced the incidence of third- or fourth-degree tears. Many found midline episiotomy associated strongly with deep tears. Some studies in the 1920s and 1930s concluded that episiotomy reduced pelvic relaxation, but they were flawed. Moreover, their conclusions do not apply to modern practice. For example, in order to prevent stretching of the pelvic floor, episiotomy would have to be performed prior to distension of the levator muscles; modern episiotomies are done much later. "[T]here is little evidence to support routine use of episiotomy. This procedure may well increase the incidence of third- and fourth-degree lacerations. There are few data to support the premise that this procedure prevents pelvic relaxation." Studies should be done to ascertain whether routine episiotomy is truly beneficial. [How much more evidence do we need that it is not?]
Episiotomies Do Not Perform as Advertised (Studies)
5. Reynolds JL and Yudkin PL. Changes in the management of labour: 2. Perineal management. Can Med Assoc J 1987;136(10):1045-1049. (mediolateral)
Vaginal births of singleton, vertex babies born between 1980 and 1984 were analyzed (N = 24,439). Changes in second stage management included more midwife-attended births, longer second stages, fewer epidurals, more spontaneous births, and, especially, fewer episiotomies. The episiotomy rate fell from 72.6% to 44.9% among primiparas and from 36.8% to 15.4% among multiparas. The percentage of intact perineums rose from 7.4% to 13.7% among primiparas and from 26.1% to 33.8% among multiparas (P < 0.001 for both). The incidence of first- and second-degree tears increased, but third-degree tears did not. [In other words, women were more likely to have a small tear, at most no worse than an episiotomy, but were not more likely to have a deep tear.] Third-degree tears were more common with forceps delivery compared with spontaneous birth (5.8% versus 2.6%, P < 0.01). Nearly all forceps deliveries included an episiotomy. Whether this is necessary has never been addressed. Rates of infection, wound breakdown, labial tears, and vulval bruising were unchanged (incidence in all cases < 0.5%). Neonatal outcomes were unchanged. Epidurals associated with episiotomy, and not having an epidural associated with intact perineum "likely due to the well-known effect of epidural analgesia on the woman's ability to push."
6. Wilcox LS et al. Episiotomy and its role in the incidence of perineal lacerations in a maternity center and a tertiary hospital obstetric service. Am J Obstet Gynecol 1989;160(5 Pt 1):1047-1052.
(both, but presume mostly midline because this is a U.S. study)
The study compared two groups of women of similar characteristics, 686 attended by midwives at a maternity center and 576 delivered by obstetric residents at a nearby hospital. No differences in incidence or type of laceration were found by episiotomy type, so results were combined.
At the maternity center, the episiotomy rate was less than that at the hospital for both primigravidas (82.7% versus 95.6%, P < 0.001) and multigravidas (44.9% versus 76.8%, P < 0.001). [These numbers are high for midwives.] After adjusting for institutional differences in other variables related to use of episiotomy, women were "less than half as likely" to have an episiotomy at the maternity center. Women with episiotomies suffered a marked increase in the probability of third-degree tear (OR 4.29 CI 1.58-11.69). Neither the number of days of maternal fever [a measure of infection] nor the incidence of birth injuries differed by use of episiotomy. Since place of delivery was a strong predictor of episiotomy, "[i]t is not unreasonable to assume that this difference . . . is related to . . . institutional differences in philosophies." Given these results, serious questions are raised about the frequent use of episiotomy.
7. Klein M et al. Does episiotomy prevent perineal trauma and pelvic floor relaxation? Online J Curr Clin Trials 1992;1(Document 10). (midline)
Between 1988 and 1990, women were randomly assigned to "restricted" (try to avoid an episiotomy) or "liberal" (try to avoid a tear) episiotomy groups. Enrollment was limited to low-risk women past 34 weeks gestation (N = 1044). Randomization took place late in second stage labor. Ultimately, 359 primiparas and 344 multiparas were randomized. Among those randomized, only 32% of primiparas and 48% of multiparas gave birth without oxytocin augmentation, epidural anesthesia, or forceps. The study had 90% power to detect a 7% difference in intact perineums among primipara and a 15% difference among multipara, assuming a 30% dropout rate.
No infant in either group required special care or had a 5-minute Apgar less than 6. Among primiparas, the episiotomy rate was reduced by one-third in the restricted group for both primiparas (57.2% versus 81.4%) and multiparas (30.7% versus 47.0%). The reasons given for episiotomy in the restricted group were severe tear anticipated (40%), fetal distress (29%), and perineum not distending (23%). "Accustomed to the liberal or routine use of episiotomy, and despite being presented with a population of healthy, low-risk women, many physicians had difficulty in withholding episiotomy in the [restricted] arm of the trial." Among primiparas, 52 (14.5%) had episiotomy extensions, of which 46 were third or fourth degree, six were tears into the upper vagina. Only one woman had a spontaneous deep tear. Among multiparas, 1.8% had third- or fourth-degree episiotomy extensions, and no one had a spontaneous deep tear. As measured by electromyographic perineometry, no differences were found between groups for pelvic floor functioning three months postpartum. Values related to parity, size of baby, and antepartum status. Episiotomy had no effect. No differences were found between groups for anterior vaginal trauma, pain, pain at resumption of intercourse, or female sexual satisfaction. Urinary incontinence was more prevalent in primiparous women in the restricted group (21.1% versus 14.5%, NS) and among multiparous women in the liberal group (12.9% versus 21.5%, NS). [Because outcomes are reported by group, instead of according to whether the woman had an episiotomy, we do not know how these findings relate to episiotomy.] Based on the results, the authors recommend that liberal or routine use of episiotomy be abandoned. This recommendation can be implemented only where the birth attendant has learned how to protect the perineum. Such training should be part of normal education and practice. At least some instrumental deliveries could be done without episiotomy, and doctors should be trained in that too.
Episiotomies Do Not Prevent Deep Tears
8. Gass MS, Dunn CD, and Stys SJ. Effect of episiotomy on the frequency of vaginal outlet lacerations. J Reprod Med 1986;31(4);240-244. (midline)
To examine the effect of episiotomy on laceration, women having spontaneous births who either had or did not have an episiotomy were matched for age (range 18-24), parity (range 0-2), and birth weight (within 200 g) (N = 205 pairs). Commonly lacerations are considered only when they occurred in addition to an episiotomy. "[This] implie[s] that an intact perineum with no episiotomy or lacerations is equivalent to a perineum with an episiotomy and no lacerations. The validity of that statement may be questioned. To the patient they are not equivalent since she must undergo the incision, incision repair and recovery." If an episiotomy is counted as a second-degree laceration, then women with no episiotomy had fewer lacerations (P < 0.0001). Analyzing lacerations this way, the woman with an episiotomy will have fewer lacerations 6 + or - 3% of the time, whereas the woman with no episiotomy will have fewer lacerations 78 + or - 5% of the time. The 2% of women who had third- or fourth-degree lacerations all had episiotomies (17% para 0, 3% para 1, 0 para 2). However, since the study was not randomized, it is possible that episiotomy was done to forestall a tear or for fetal distress, which might raise the third- and fourth-degree tear rate. "The results of this study do not support routine use of episiotomy."
9. Thorp JM et al. Selected use of midline episiotomy: effect on perineal trauma. Obstet Gynecol 1987;70(2):260-262. (midline)
One author, attending 113 women, limited his use of episiotomy to instrumental deliveries or when there was fetal distress. His episiotomy rate was 14%. He compared his outcomes with the rest of the residents, who, attending 265 women, had an unrestricted policy. Their episiotomy rate was 63%. Among all women, the incidence of third- or fourth degree tears in the limited group was 13.2% versus 1.8% (P <+0.05). Both nulliparas (22.1% versus 1.6%, P < than or = to 0.001) and multiparas (5.6% versus 1.8%, NS) were less likely to have deep tears in the restricted group. All third- and fourth- degree lacerations were preceded by an episiotomy. Deep tear rates agree with other published reports. In no case was there injury to the urethra or bladder, and there were no differences in periclitoral or periurethral tears. The only difference in perineal trauma in the restricted group was that nulliparas were more likely to have a first-degree tear. All lacerations that occurred in the absence of episiotomy were easy to repair.
10. Green JR and Soohoo SL. Factors associated with rectal injury in spontaneous deliveries. Obstet Gynecol 1989;73(5 Pt 1):732-738. (midline)
Logistic regression was done on 2706 spontaneous cephalic births to determine factors associated with rectal injury (third and fourth degree). The rectal injury rate was 13.0%. Ninety percent occurred with episiotomy. The episiotomy rate was 45.5% among doctors versus 27.4% among midwives (P < 0.0001). The intact perineum rate among women who did not have an episiotomy (N = 760) was 47.7%. The adjusted odds ratio for rectal injury was 8.9 for episiotomy versus no episiotomy, 3.3 for nulliparity versus multiparity, 2.4 for physician versus midwife, and 2.0 for delivery room birth versus a labor bed. (Birth in the delivery room was in the lithotomy position; in the labor bed, a variety of positions were used, with semirecumbent the most common.) "It seems likely that midwives have the patience to allow for slow stretching of the perineum . . . and this may be an important factor that distinguishes them from many physicians." Although cause and effect cannot be proved in an observational study, the magnitude of the difference makes noncausal explanations unlikely.
11. Shiono P, Klebanoff MA, and Carey JC. Midline episiotomies: more harm than good? Obstet Gynecol 1990;75(5):765-770. (both)
The association between episiotomy and severe (third and fourth degree) lacerations was studied in 24,114 women having a singleton, vertex birth of a baby weighing over 500 g. The severe tear rate was 8.3% among primiparas and 1.5% among multiparas. Women with midline episiotomies were 50 times more likely to have a severe tear (OR 12.5 among primiparas; OR 32.3 among multiparas), and women with mediolateral episiotomies were eight times more likely to have a severe tear (OR 1.2 among primiparas; OR 5.3 among multiparas) compared with women with no episiotomy. Forceps increased the odds eightfold. All differences were significant. After statistical adjustment for confounding factors, midline episiotomy increased the risk of deep tears 4.2-fold among primiparas and 12.8-fold among multiparas. Mediolateral episiotomy reduced the risk of deep tears 2.5-fold among primiparas and increased it 2.4-fold among multiparas (NS) compared with no episiotomy. Forceps use tripled the probability of severe lacerations. Although mediolateral episiotomy offered some protection against severe tears for primiparas, studies have found they have other problems, including increased pain, poorer cosmetic results, and more dyspareunia. Since most rectovaginal fistulas occur after severe tears, reducing the rate of these tears would result in fewer fistulas and less morbidity.
Episiotomies Do Not Heal Better than Tears
12. McGuiness M, Norr K, and Nacion K. Comparison between different perineal outcomes on tissue healing. J Nurse Midwifery 1991;36(3):192-198. (both but presumably mostly midline because this is a U.S. study)
The article begins with a brief literature review concluding that episiotomy does not improve healing, prevent pelvic floor relaxation, prevent lacerations, or protect the fetus. Episiotomy increases the risk of infection, including fatal infection, causes short-term and long-term pain and dyspareunia, and causes "considerable" blood loss. Perineal healing was compared one to two weeks postpartum between 181 women with episiotomies and 186 women without them (episiotomy rate 49%). All women were indigent low-risk women who had spontaneous births. Only 2% of the no-episiotomy group had third-degree lacerations versus 15% of the episiotomy group (third- and fourth-degree tears were collapsed into one category). In the episiotomy group 7.7% of women experienced delayed healing versus 2.2% in the no-episiotomy group (P < 0.05). The difference persisted when women with intact perineums (53% of the no-episiotomy group) were excluded. None of the four third-degree lacerations that occurred without episiotomy exhibited delayed healing versus 18.5% of the 27 deep tears that occurred with epsiotomy. Both infections occurred in women with epsiotomies (midline with first-degree periurethral tear; mediolateral with third-degree extension). Women who had instrumental deliveries were excluded. However, this group had a 35% third-degree extension rate, and delayed healing occurred in 17.6%, confirming that instrumental delivery increases perineal trauma. "This suggests that women without episiotomies exhibit better perineal healing than women with episiotomies."
Episiotomies Do Not Prevent Pelvic Floor Muscle Relaxation
13. Gordon H and Logue M. Perineal muscle function after childbirth. Lancet 1985;2:123-125. (unstated, but presumably mediolateral because this is a British study)
The strength of the perineal muscles was measured in primiparas one year after delivery. Women with an intact perineum, a second-degree laceration, an episiotomy, and a forceps delivery with episiotomy were compared with women who had a cesarean and nulliparas. No differences were found. Women were also grouped by whether they exercised regularly, had done only postpartum exercises, or never exercised. Regardless of perineal history, women who exercised regularly fared best, those who did postpartum exercises fell in the middle, and those who did not exercise did worst (P < 0.0001 for no exercise compared with postpartum exercise and regular exercise). Exercise regimens included fitness classes, walking, jogging, running, swimming, dancing, and yoga. Few women specifically did pelvic floor contractions [Kegels]. "This study does not support the theory that episiotomy results in improved healing and better perineal muscle function."
14. Sleep J et al. West Berkshire perineal management trial. Br Med J;289:587-590. (mediolateral)
Women were allocated to either "avoid episiotomy" [restricted group] (N = 498) or "prevent a tear" [liberal group] (N = 502). The study had 90% power to detect a significant difference if restricted policy doubled the incidence of an outcome expected in 5% of cases.
The episiotomy rate was 10% in the restricted group versus 51% in the liberal group. The restricted group experienced more perineal tears and labial tears and had more intact perineums. "Severe maternal trauma" (defined as extension through the anal sphincter, the rectal mucosa, or to the upper third of the vagina) occurred in four cases in the restricted group versus one case in the liberal group (NS). More women in the liberal group required suturing (78% versus 69%, P < 0.01). Neonatal outcomes were similar. There were no differences in reporting of pain [which may reflect only that pain relief measures were effective] or in the use of analgesics at 10 days postpartum (3% versus 2%). Twelve percent of each group sought medical advice for perineal problems. However, 37% of the restricted group versus 27% of the liberal group resumed sexual intercourse within one month after birth (P < 0.01). By three months postpartum, the prevalence of dyspareunia was similar (22% versus 18%). Equal numbers of women were breastfeeding at 10 days (70%) and 3 months (48%). [This is a confounding factor for dyspareunia. Breastfeeding women may experience vaginal dryness due to low estrogen levels.] No differences were found for stress incontinence. "A large proportion of women (19%) had involuntary loss of urine three months after delivery, but there is no evidence from this study that the liberal use of episiotomy prevents this problem."
15. Sleep J and Grant A. West Berkshire perineal management trial: Three year follow up. Br Med J 1987;32(3):181-183. (mediolateral)
Of the original 1000 women [see Abstract 14], 674 filled out a questionnaire three years later. The response rate was 91% among those still at the same address. Dyspareunia was reported by 16% of the restrictive episiotomy group versus 13% of the liberal group. Stress incontinence was reported among 34% of the restrictive versus 36% of the liberal group. For 9% and 8%, respectively, problems were severe enough to wear a pad. Three of the four women in the restrictive group with severe trauma had minor degrees of urinary incontinence and/or dyspareunia. The one woman with severe trauma in the liberal group reported no problems. "There was no clear difference between the groups with respect to dyspareunia. There was also no difference in the prevalence of urinary incontinence, even when the severity and nature of the incontinence, and subsequent deliveries, were taken into account."
16. Rockner G, Jonasson A, and Olund A. The effect of mediolateral episiotomy at delivery on pelvic floor muscle strength evaluated with vaginal cones. Acta Obstet Gynecol Scand 1991;70(1):51-54. (mediolateral)
Pelvic floor muscle strength pre- and postpartum was evaluated by measuring the mother's ability to retain vaginal cones of various weights for one minute while upright. The population was 87 primigravidas, of whom 16 had elective c-section. Among vaginal births, 30% had episiotomy, 36% had spontaneous tears, and 34% had intact perineums. Two women had third-degree tears in the episiotomy group versus one in the spontaneous tear group. Compared with spontaneous tear (P < 0.001) and intact perineum (P < 0.01), pelvic floor strength was weakest in the episiotomy subgroup (mean decrease in muscle strength 30.0 + or - 11.8). Those with spontaneous tears (mean decrease 18.9 + or - 9.1) were similar to intact perineum (mean decrease 19.2 + or - 10.2). Those who had elective cesareans (mean decrease 0) faired the best compared with all subgroups of vaginal birth (P < 0.01). [Median episiotomy may give different results compared with mediolateral episiotomy because fewer muscle groups are cut. In any case, spontaneous laceration and intact perineum were similar, which means not doing an episiotomy does no harm and that the major effect on pelvic floor function may be intrinsic to vaginal birth.] "[T]he present results do not support the concept that [episiotomy] reduces the risk of damage to pelvic floor muscles.]
Episiotomies Do Not Prevent Fetal Brain Damage
Note: If episiotomy has no protective effect in low- and very-low-birth-weight babies, a population highly vulnerable to brain damage from stress or trauma, benefits for a full-term, healthy baby seem unlikely.
17. Lobb MO, Duthie SJ, and Cooke RW. The influence of episiotomy on the neonatal survival and incidence of periventricular haemorrhage in very-low-birth-weight infants. Eur J Obstet Gynecol Reprod Biol 1986;22(1-2):17-21.
The protective effect of episiotomy on infants weighing 1500 g or less and free of lethal abnormalities was examined. Some doctors routinely did episiotomy in such cases (N = 43 infants), others did not (N = 51 infants). When all very-low-birth-weight (VLBW) babies were compared, episiotomy appeared to improve survival rates and decrease incidence of periventricular hemorrhage. "However, . . . when VLBW babies of similar weight and age are considered, the use of episiotomy appears to hold no advantages."
18. The TG. Is routine episiotomy beneficial in the low birth weight delivery? Int J Gynaecol Obstet 1990;31(2):135-140.
The protective effect of episiotomy was analyzed for 439 singleton, vertex, spontaneous births of babies weighing less than 2500 g. Preexisting pregnancy complications were excluded. Since episiotomy rate depends on parity, primiparas and multiparas were analyzed separately. Neonatal mortality rates were similar for episiotomy versus no episiotomy among both nulliparas (6.7% versus 8.6%) and multiparas (10.0% versus 9.3%) as were 5-minute Apgar scores. One of the two postulated protective effects of episiotomy is reduced pressure on the fetal head. However, to achieve this, episiotomy would have to be done before the head distended the perineum. The other protective effect is shortening second stage. However, length of second stage does not correlate consistently with outcome. "The results of this . . . study suggest that routine episiotomy has little, if any beneficial effect on . . . neonatal outcome."