- Email: [email protected]
Incidence of postpartum lumbosacral spine and lower extremity nerve injuries
Incidence of Postpartum Lumbosacral Spine and Lower Extremity Nerve Injuries Cynthia A. Wong, MD, Barbara M. Scavone, MD, Sheila Dugan, MD, Joanne C. Smith, MD, Heidi Prather, DO, Jeanne N. Ganchiff, MPH, and Robert J. McCarthy, PharmD OBJECTIVE: Neurological injury associated with present day labor and delivery is thought to be unusual. The purpose of this study was to estimate the incidence, severity, and duration of postpartum lumbosacral spine and lower extremity nerve injury and identify factors related to nerve injury. METHODS: All women who delivered a live-born infant from July 1997 through June 1998 were asked about symptoms of lumbosacral spine and lower extremity nerve injury the day after delivery. Women with symptoms were examined by a physiatrist to confirm injury, and their cases were then followed by telephone until the symptoms resolved. Maternal variables (including prospective documentation of time spent pushing in various positions) and fetal variables that might be associated with risk of nerve injury were compared between women with injury and those without. RESULTS: Six thousand fifty-seven women delivered liveborn infants; 6048 were interviewed and 56 had a confirmed new nerve injury, an incidence of 0.92%. Factors found by logistic regression analysis to be associated with nerve injury were nulliparity and prolonged second stage of labor. Women with nerve injury spent more time pushing in the semi-Fowler–lithotomy position than women without injury. The median duration of symptoms was 2 months. CONCLUSION: The estimated incidence of postpartum nerve injury was greater than reported from previous studies and is associated with nulliparity and prolonged second stage of labor. (Obstet Gynecol 2003;101:279 – 88. © 2003 by The American College of Obstetricians and Gynecologists.)
The reported incidence of postpartum lower extremity motor and sensory dysfunction secondary to neurological injury in present day obstetric practice is between 0.008% and 0.5%.1–5 The incidence has been inversely From the Departments of Anesthesiology and Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University, Chicago, Illinois. SD is currently affiliated with Rush Medical College, Chicago, Illinois, and HP with Washington University School of Medicine, St. Louis, Missouri.
related to sample size and varies widely with study methodology. In a retrospective review of 23,827 deliveries over a 9-year period the incidence of paresthesias and motor dysfunction was found to be 18.9 per 10,000 deliveries.1 A second retrospective review of 143,019 live births over a 16-year period reported an incidence of 0.8 per 10,000 births.2 Similarly, two prospective physician-reported audits from the early 1990s of 467,491 and 48,066 deliveries identified nerve injury rates of one per 10,000 and four per 10,000 respectively.3,4 In contrast, in a prospective, case-controlled study of 3341 parturients who received regional analgesia or anesthesia for labor and delivery, symptoms of nerve injury were reported at a rate of 58 per 10,000 deliveries.5 Anecdotal experience at our institution, based on neurological symptoms uncovered during routine postanesthetic visits, indicated that the incidence of nerve injury was higher than that estimated in previous studies. We hypothesized that the true incidence of postpartum nerve injury was higher than that reported in previous studies. The purpose of this study was to prospectively estimate the incidence of postpartum lumbosacral spine and lower extremity nerve injury in parturients receiving present day obstetric care, to assess the severity and duration of injury, and to determine factors associated with injury. MATERIALS AND METHODS The study was approved by the Institutional Review Board and spanned a 1-year period from July 1997 to June 1998. The calendar day after delivery (8 –32 hours after delivery), every parturient who delivered a liveborn infant at Prentice Women’s Hospital was asked a standard question by one of the investigators: “Do you have any leg numbness or weakness?” Spanish-speaking women were asked to read the question on a card. Interpreters were used for other languages. If the answer was affirmative or equivocal, the investigator performed a preliminary neurological examination to ascertain that neurological pathology was probably present. She then
VOL. 101, NO. 2, FEBRUARY 2003 © 2003 by The American College of Obstetricians and Gynecologists. Published by Elsevier.
0029-7844/03/$30.00 doi:10.1016/S0029-7844(02)02727-8
279
asked the woman for informed, written consent to participate in study follow-up, which consisted of a complete lower extremity neurological examination by a physiatrist investigator within 24 hours of interview and further diagnostic workup and referrals for physical therapy and rehabilitation, as clinically indicated. Thereafter, women with confirmed nerve injury (as diagnosed by physical examination performed by a physiatrist) were contacted by phone by a research nurse every week for 4 weeks, and then every month until symptom resolution. Women who answered in the affirmative regarding leg numbness or weakness but who refused follow-up were considered to have suspected but not confirmed nerve injury. The following information was documented on admission (by the nursing staff) or immediately postpartum (by the nursing staff, delivering obstetrician, or midwife) and was abstracted from an institutional database for analysis: maternal age, height, weight, prepregnancy weight and pregnancy weight gain, parity, type of labor analgesia, duration of second stage of labor, fetal presentation (vertex or nonvertex), type of delivery (spontaneous, vacuum- or forceps-assisted vaginal, or elective or scheduled versus nonscheduled cesarean delivery), and newborn weight. During the study period, the nursing staff documented fetal station and position at the time maternal pushing was initiated, and duration of pushing. Pushing position was documented in epochs of 15 minutes duration with the primary position during the epoch (lateral, supine, Fowler–lithotomy with thigh flexion less than 90° or with thigh flexion greater than 90°, pushing stool, pushing bar, squatting, or kneeling). Pregnancy weight gain was classified as low, normal, or high as defined by the National Academy of Science Institute of Medicine criteria.6 Prolonged second stage of labor was defined using The American College of Obstetricians and Gynecologist’s criteria, depending on parity and the presence or absence of neuraxial labor analgesia.7 Data were compared between women with elective or scheduled cesarean delivery and those with all other deliveries by 2, the two-sample t test, or the two-tailed Kolmogorov–Smirnov two-sample test. The proportion of women reporting symptoms of nerve injury were compared between women with elective or scheduled cesarean delivery and those with all other methods using a 2 statistic. In women who labored, univariable analysis of parturient and fetal characteristics believed to be associated with the potential for nerve injury were compared between women with confirmed nerve injury and those without by the 2 or Fisher exact test (parity [nulliparous versus parous], regional analgesia, fetal position [vertex versus nonvertex], prolonged second stage
280
Wong et al
Postpartum Nerve Injury
of labor, pregnancy weight gain [low, normal, high], prepregnancy body mass index [BMI] [underweight, normal, overweight6], infant weight [greater than 50th percentile], method of delivery, and gestational age [term versus preterm]) or the two-sample t test or the two-tailed Kolmogorov–Smirnov two-sample test (maternal age, height, weight, and BMI at time of delivery). A P value of less than .05 was required to reject the null hypothesis. Logistic regression analysis was performed using a backward-stepwise likelihood ratio elimination method with maternal and fetal variables for which P ⬍ .2 on univariable analysis. Models were developed for confirmed nerve injuries, and confirmed plus suspected nerve injuries. Criteria for entry and removal from the equation were set at .05 and .1, respectively. Concordance of the model with the observed outcomes was determined by measuring the area under the receiver operating characteristic curve. Bivariate correlation between duration of pushing and duration of the second stage of labor was determined using the Spearman . Box plots of the number of 15-minute epochs spent pushing in semi-Fowler–lithotomy less than 90°, greater than 90°, and other positions (identifying median and interquartile range, tenth to 90th percentile range, and mean) were constructed to compare women with nerve injury and those without nerve injury. The number of epochs in each pushing position in parturients with and without confirmed nerve injury was compared using the Mann–Whitney U test. Kaplan–Meier survival curves (percentage of women in each group with continuing symptoms of nerve injury) were constructed, and the median duration of symptoms was determined. The log rank test was used to compare symptom duration in the femoral nerve, lateral femoral cutaneous, and other nerve distributions, and to compare duration of symptoms involving sensory-only versus sensory–motor deficits. RESULTS Six thousand fifty-seven mothers delivered live-born infants during the study period. The demographic, labor, and fetal characteristics of these 6057 women are listed in Table 1. There were multiple differences between women who had an elective or scheduled cesarean delivery and those who did not. Nine women were not queried because of a language barrier or early discharge from the hospital, but were assumed not to have nerve injury for purposes of data analysis. The incidence of nerve injury is delineated in Table 2. Women with nerve injuries that were preexisting (by history) or not confirmed by physiatrist’s examination were considered without nerve injury. The incidence of
OBSTETRICS & GYNECOLOGY
Table 1. Parturient Characteristics
Maternal Age (y) Height (in.) Prepregnancy weight (kg) Prepregnancy BMI (kg · m⫺2) Weight gain (kg) Delivery weight (kg) Nulliparous (%) Gestational age (wk) Labor analgesia None (%) Systemic only (%) Regional only (%) Systemic and regional (%) Infant presentation* Cephalad OA (%) Cephalad OP/OT (%) Transverse lie (%) Breech (%) Other (%) Weight (g) Multiple gestation (%)
Labor (n ⫽ 5603)
No labor (n ⫽ 454)
P
30.2 ⫾ 5.9 65 ⫾ 3 63 ⫾ 13 23.5 ⫾ 4.8 14 ⫾ 5 78 ⫾ 14 51.7 39.0 ⫾ 2.2
32.5 ⫾ 5.7 65 ⫾ 3 67 ⫾ 18 25.0 ⫾ 6.6 14 ⫾ 5 81 ⫾ 17 43.0 37.8 ⫾ 2.8
.001 .710 .001 .001 .540 .001 .001 .001
13 15 55 17
NA NA NA NA
90.7 7.5 0.2 1.4 0.2 3347 ⫾ 598 2.3
58.3 0.9 3.8 36.2 0.9 3188 ⫾ 799 9.0
}
.001 .001 .001
BMI ⫽ body mass index; OA ⫽ occiput anterior; OP ⫽ occiput posterior; OT ⫽ occiput transverse. * Immediately before delivery.
confirmed nerve injury for women undergoing elective or scheduled cesarean delivery (0.22%) did not differ from that for all other women (0.98%) (P ⫽ .125, Fisher exact test). The relative risk of nerve injury for laboring women versus nonlaboring women in this study was 0.99 (95% confidence interval 0.99, 1.00). The distribution of nerve injuries is listed in Table 3. All diagnoses were made by physical examination. Lateral femoral cutaneous neuropathies, followed by femoral neuropathies, were the most common findings. Radiculopathies were at the L4, L5, and S1 dermatome levels. Overall, 21 women had a motor deficit, an incidence of 0.37%. Univariable analysis comparing laboring women with confirmed nerve injury to laboring women without nerve injury showed that laboring women with postparTable 2. Incidence of Nerve Injuries
Eligible participants (N ⫽ 6057) Interviewed Symptoms on interview Examined by physiatrist Findings [n (%)] Preexisting Not confirmed New confirmed
VOL. 101, NO. 2, FEBRUARY 2003
Labor [n (%)]
No labor [n (%)]
5603 (92.5) 5594 (99.9) 80 (1.4) 72 (90)
454 (7.5) 454 (100) 4 (0.9) 4 (100)
9 (12.5) 8 (11) 55 (76.5)
1 (25) 2 (50) 1 (25)
tum nerve injury were more likely to be nulliparous, have prolonged second stage of labor, and have assisted (forceps or vacuum) vaginal deliveries (Table 4). There were no differences between laboring women with nerve injury and those without injury in maternal age, height, weight, prepregnancy BMI, pregnancy weight gain, gestational age, fetal presentation, use of regional labor analgesia, or newborn weight. Backward stepwise logistic regression analysis identified nulliparity and prolonged second stage of labor as factors associated with confirmed nerve injury (Table 5). Concordance of the model as determined by the area under the receiver operating characteristic curve was 0.65. When the eight women with suspected nerve injury who were not examined by a physiatrist (because of discharge from the hospital before an examination [n ⫽ 7] or refusal to participate in the study [n ⫽ 1]) were analyzed with the confirmed injuries, regional labor analgesia was included in the model (Table 5). Concordance of this model with observed outcome was 0.67. Pushing data (pushing time and position, fetal station at initiation of pushing) were available for 70% of parturients without injury and 91% of parturients with nerve injury. The position of the fetal presenting part at initiation of pushing was not documented for the majority of parturients; therefore the available data were not analyzed. Little time was spent in pushing positions other
Wong et al
Postpartum Nerve Injury
281
Table 3. Distribution of Nerve Injuries in Laboring Parturients New Lateral femoral cutaneous Unilateral Bilateral Femoral (sensory deficit only) Unilateral Femoral (sensory and motor deficit) Unilateral Bilateral Common peroneal† Unilateral Lumbosacral plexus Unilateral Sciatic Unilateral Obturator Unilateral Radicular Unilateral Other Unilateral Totals
Total new
Confirmed
Suspected*
Preexisting
22 2
21 2
1 0
8 0
8
5
3
0
13 1
13 1
0 0
0 0
3
3
0
0
3
3
0
0
2
1
1
1
3
0
3
0
5
5
0
0
1 63
1 55
0 8
1 10
* Women reporting new symptoms of nerve injury who were not examined by a physiatrist. † Also known as the fibular nerve.
than semi-Fowler–lithotomy less than 90° thigh flexion and semi-Fowler–lithotomy greater than 90° thigh flexion; therefore these positions were combined into an “other” category for purposes of data analysis. There was a difference in the number of 15-minute epoch parturients with nerve injuries spent in either semi-Fowler–lithotomy position versus parturients with no nerve injury (Figure 1). The mean difference in pushing time and duration of second stage was 9 ⫾ 36 minutes, and pushing time was highly correlated with the duration of the second stage of labor ( ⫽ 0.912, P ⬍ .01). Women with nerve injuries began pushing with the fetal presenting part at a significantly higher station (P ⫽ .03) than women without nerve injury (Figure 1), and fetal station was correlated with both pushing time ( ⫽ ⫺0.3, P ⬍ .01) and duration of the second stage of labor ( ⫽ ⫺0.3, P ⬍ .001). The median duration of nerve injury symptoms was 2 months (range 1 week to greater than 18 months) (Figure 2). There was no difference (log rank 0.76, P ⫽ .68) in the median duration of symptoms when compared by nerve distribution (Figure 2, inset), or sensory-only injuries compared with sensory–motor (log rank 0.01, P ⫽ .93). Nine women were lost to follow-up (lateral femoral cutaneous nerve [three], femoral nerve [two], and other [four]). Two women continued to have symptoms (occasional pain) after 14 and 18 months, respectively. Six of the eight patients with suspected nerve injury received
282
Wong et al
Postpartum Nerve Injury
follow-up phone calls, with the duration of symptoms lasting between 2 weeks and 18 months (median 2 months). DISCUSSION Neurological injury associated with pregnancy and delivery, or intrinsic obstetric palsy, has long been recognized. The incidence of femoral neuropathy around 1900 was 3.2%, and 25% of cases were bilateral.8 Historically, case reports of postpartum nerve injury were associated with primiparity, cephalopelvic disproportion, and midforceps deliveries.8 The results of the current study support historical experience, in that nulliparas with prolonged second stage of labor were more likely to experience postpartum nerve injury, and that the majority of these nerve injuries were intrinsic obstetric palsies and were not directly related to regional anesthesia or analgesia.8 The incidence of postpartum injury in the current study was 1%, higher than the estimated incidence from retrospective studies performed between the late 1970s and early 1990s,1–3 but on the same order of magnitude as a recent prospective study in which postpartum women who received regional analgesia or anesthesia were asked a question similar to that in the current study (0.57%).5 There may be several reasons for the large difference in incidences found in previous studies per-
OBSTETRICS & GYNECOLOGY
Table 4. Univariable Analysis of Risk Factors in Laboring Parturients
Prolonged second stage (%) No Yes Nulliparity (%) No Yes Delivery (%) NSVD Vaginal, assisted Cesarean Regional labor analgesia (%) No Yes Newborn weight ⬎ 50th percentile (%) No Yes Prepregnancy BMI (kg 䡠 m⫺2) (%)† ⬍19.8 ⱖ19.8 and ⱕ26.0 ⬎26.0 Presentation/position (%) Cephalad, OA Cephalad, OP Other Gestation (%) ⬍37 wk ⱖ37 wk Weight gain (%)‡ Low Normal High Height (in.) Current weight (kg) Age (y)
No nerve injury (n ⫽ 5548)*
Confirmed nerve injury (n ⫽ 55)
90.9 8.1
76.4 23.2
.001
48.6 51.4
25.5 74.5
.001
81.7 9.6 8.7
69.1 20.0 10.9
28.0 72.0
16.4 83.6
.055
54.2 45.8
43.6 56.4
.116
18.4 60.0 21.6
27.8 48.1 24.1
.136
90.5 7.7 1.8
83.6 14.6 1.8
.163
8.4 91.6
3.6 96.4
23.7 38.6 37.7 64.6 ⫾ 2.8 77 ⫾ 14 30 ⫾ 6
23.4 48.9 27.7 64.3 ⫾ 2.7 79 ⫾ 18 30 ⫾ 6
P
.025
.205
.283 .375 .461 .778
NSVD ⫽ normal spontaneous vaginal delivery; vaginal assisted ⫽ forceps- or vacuum-assisted vaginal delivery. Other abbreviations as in Table 1. * Includes women with symptoms on interview not examined by physiatrist. † Underweight, normal, or overweight as defined by the Institute of Medicine.6 ‡ As defined by the Institute of Medicine.6
formed in the present era of obstetric care and that in the current study. The reported incidence of nerve injury 2 decades ago may have been falsely low.5,9 Retrospective
studies have known limitations. For example, none of the women diagnosed with nerve injury in our study were coded for such a diagnosis in their medical record.
Table 5. Multivariable Predictors of Nerve Injury in Laboring Parturients Variable Nulliparity Prolonged labor Constant Nulliparity Prolonged labor Regional labor analgesia Constant

SE
P
Exp ()
Women with confirmed nerve injury 0.94 0.34 .005 2.55 0.84 0.38 .025 2.32 ⫺5.41 0.29 .000 0.004 Women with suspected and confirmed nerve injury 0.95 0.30 .002 2.57 1.07 0.31 .001 2.92 0.69 0.36 .057 2.00 ⫺5.80 0.40 .000 0.003
95% CI for Exp () Lower
Upper
1.32 1.11
4.94 4.84
1.43 1.59 0.98
4.64 5.36 4.08
SE ⫽ standard error.
VOL. 101, NO. 2, FEBRUARY 2003
Wong et al
Postpartum Nerve Injury
283
Figure 1. Box plot of the number of 15-min epochs spent pushing in semi-Fowler–lithotomy less than 90°, greater than 90°, and other positions, and fetal station at initiation of pushing, for women with nerve injury versus those without nerve injury. Median shown as solid line and mean as dotted line, with box representing 25th to 75th percentile and whiskers tenth and 90th percentiles, respectively. Data are compared between groups using the Mann–Whitney U test. †P ⬍ .05. Wong. Postpartum Nerve Injury. Obstet Gynecol 2003.
Therefore, a retrospective search by diagnosis code2 would have yielded an incidence of zero. Vargo postulated that a change in obstetric management (increase in cesarean delivery rate, decrease in the use of midrotational forceps, and a decrease in the duration of the second stage of labor) contributed to a reduced incidence of nerve injury during the 1970s and 1980s relative to reports from earlier in the 20th century. The national cesarean delivery rate increased markedly during the 1970s and 1980s (17.5% in 1988); however, it declined during the 1990s,10 particularly at our institution, where the primary cesarean delivery rate at the time of the study was 10.8%. During this same period the national rate of regional labor analgesia in large hospitals increased from 16% in 1981 to 66% in 199711 and was 72% in our institution at the time of this study. Although regional labor analgesia was not directly associated with confirmed nerve injury in the current study, a large proportion of women received it, and therefore a larger study with more women not receiving regional analgesia would be necessary to adequately address whether regional labor analgesia is associated with postpartum nerve injury. Regional labor analgesia is associated with a longer
284
Wong et al
Postpartum Nerve Injury
second stage of labor,12 and parturients with labor analgesia–induced sensory blockade may not appreciate symptoms of impending nerve injury13 and may not change their body position in a timely manner. The presence of regional analgesia may encourage parturients to push without changing body position frequently, or may have contributed to an overall change in the proportion of time parturients push in various positions. Our data do not allow us to differentiate whether prolonged pushing (with flexion and external rotation of the thighs) or prolonged second stage is associated with nerve injury because the duration of pushing in this study was highly correlated with the duration of the second stage. Most parturients at our institution are instructed to push once the cervix is completely dilated. Acute injury to nerves can occur by several mechanisms, including transection, stretch, compression, or vascular injury.14 Compression and stretch may result in decreased perineural blood flow and ischemia. Less severe injuries cause focal demyelination and conduction block, whereas more severe injuries cause axon loss resulting in nerve conduction failure.15 Recovery of nerve function is best predicted by the percentage of axonal loss, with axonal loss less than 50% predicting
OBSTETRICS & GYNECOLOGY
Figure 2. Kaplan–Meier plot of the duration of neurological symptoms in women with confirmed new nerve injuries. Inset panel shows duration of symptoms by nerve distribution. Wong. Postpartum Nerve Injury. Obstet Gynecol 2003.
recovery within 1 year.16 Based on the duration of their symptoms, the women in this study likely suffered minor degrees of axon loss or focal demyelination. Interestingly, Ong and colleagues1 found that symptoms of nerve injury had resolved within 72 hours in the 45 women they identified, compared with a median duration of symptoms of 2 months in the current study. Others have found that symptoms usually resolve in less than 6 months.5 The majority of women had injury to the lateral femoral cutaneous nerve, also known as meralgia paresthetica. Risk factors for lateral femoral cutaneous neuropathy include anatomic variation (the nerve may bisect the inguinal ligament), increased intraabdominal pressure, pregnancy, obesity, diabetes mellitus, trauma, pressure by a belt, or prolonged hip flexion.17 Exacerbated lumbar lordosis (eg, during pregnancy) leads to compression of the nerve by the posterior fascicle of the inguinal ligament in patients in whom the nerve bisects the inguinal ligament.18 Indeed, nine parturients identified in this study described symptoms consistent with this syndrome before delivery. The nerve may be entrapped or compressed by the inguinal ligament during maternal pushing in the thigh-flexed position. Interestingly, four women with this neuropathy underwent a
VOL. 101, NO. 2, FEBRUARY 2003
cesarean delivery before the second stage of labor. The nerve may have been injured during cesarean delivery, by cutting (a wide Pfannenstiel incision), stretching, or pressure by a retractor.19 Alternatively, laboring parturients usually wear one to two elastic belts around their lower abdomen to hold monitors in place. Perhaps these belts or monitors applied pressure on the nerve. The second most common injury was to the femoral nerve. Injury to the femoral nerve has been described after lower abdominal surgery using a Pfannenstiel incision, particularly with self-retaining retractors.20 However, no women with femoral nerve injury in our study had a cesarean delivery. The femoral nerve does not descend into the true pelvis, and therefore it is unlikely that compression by the fetal head or body is a mechanism of injury. The nerve may be compressed at the inguinal ligament during thigh flexion, external rotation, and abduction. However, if weakness of thigh flexion is present (iliopsoas muscles), then the lesion must be proximal to the inguinal ligament, as nerve fibers to the iliopsoas muscle branch proximal to the inguinal ligament.21 All 13 women in the current study with a motor component to their femoral neuropathy had weakness of hip flexion as well as knee extension, thereby placing the likely site of injury proximal to the inguinal ligament.
Wong et al
Postpartum Nerve Injury
285
Femoral nerve injury with iliopsoas weakness has been described after vaginal hysterectomies in the lithotomy position, and cadaver studies showed that retractor replacement was not responsible for nerve injury.21 The intrapelvic portion of the femoral nerve has a poor blood supply relative to other portions of the nerve.21 Several authors have suggested that excessive hip abduction and external rotation may cause a stretch lesion of the intrapelvic portion of the nerve22 and that poor blood supply may contribute to stretch-induced nerve ischemia.9 Finally, variant slips of the psoas and iliacus muscles that split the femoral nerve have been identified. They may place tension on the femoral nerve and lead to an entrapment syndrome.23 Five laboring women presented with history and physical examination findings most consistent with injuries at the radicular level. Radicular symptoms can be caused by many processes, including bulging or ruptured intervertebral disks and, less commonly, arachnoiditis, spinal cord, intra- or extradural tumors, trauma, infection, inflammation, or even muscle spasm.17 There is contradictory evidence as to whether pregnancy and labor predispose to intervertebral disk prolapse and subsequent radiculopathy.24 Although all five women in the current study had regional labor analgesia (three had combined spinal– epidural analgesia, and two had epidural analgesia), no women had paresthesias or pain during epidural or spinal needle insertion or drug injection. In contrast to the laboring patients diagnosed with radiculopathy, the single woman who had a nerve injury after elective cesarean delivery had postpartum radicular pain in the same distribution as a paresthesia elicited during initiation of spinal anesthesia. Five days after delivery her symptoms were improving, but she was lost to further follow-up. Recent prospective studies in obstetric patients have found a similarly low incidence of traumatic neuropathy (one in 10,99525 and one in 13,0074). In the middle of the 20th century foot drop was the most common obstetric nerve palsy.8 Foot drop can be secondary to injury at the level of the lumbosacral root, lumbosacral plexus, sciatic nerve, or the common peroneal nerve (also known as the fibular nerve). Even if the injury occurs at the level of the lumbosacral plexus, the common peroneal nerve component is more often affected than the tibial component of the plexus because of its relationship to the bony pelvis.18 Components of the lumbosacral plexus cross the pelvic brim or originate in the true pelvis and could be compressed between the bony pelvis and fetal head (or forceps). Indeed, fetal macrosomia; malpresentations such as occiput posterior, or brow; or specific pelvic features such as a straight
286
Wong et al
Postpartum Nerve Injury
sacrum or wide, posterior pelvis are described as risk factors for obstetric-related foot drop.18 The most common mechanism of peroneal nerve injury is external compression15 classically from pressure exerted by inappropriate leg positioning in stirrups. Two case reports describe postpartum peroneal neuropathy in women with prolonged pushing with the knees hyperflexed and the parturients’ hands grabbing and applying pressure over the lateral, upper leg.26,27 The nerve may also be injured during prolonged squatting during labor.28,29 Injury to the obturator nerve has also been described in the postpartum period. The nerve crosses the pelvic brim and may be compressed between the pelvis and fetal head or forceps applied to the fetal head.8,18 The lithotomy position worsens angulation of the nerve as it leaves the obturator foramen.18 Pudendal nerve blocks have caused hematomas that have led to entrapment neuropathies of the obturator nerve.8 Similar to this study, postpartum injury to the sciatic nerve has been reported,30 although the mechanism remains unclear.18,30 Stretch injuries to the sciatic nerve after gynecologic procedures in the lithotomy position have been described.18 In a prospective study of lower extremity neuropathies associated with 991 surgical procedures performed in the lithotomy position,31 the incidence of nerve injury was 1.5% and included obturator, lateral femoral cutaneous, sciatic, and peroneal nerve injuries. In agreement with our findings, nerve injury was associated with time in the lithotomy position, but was not associated with patient body habitus or diabetes mellitus. Symptoms resolved in 93% of patients by 6 months. There are several limitations to our study. The current study may have underestimated the incidence of lower extremity nerve injuries for a number of reasons. Some women may not have reported symptoms of nerve injury when interviewed, and women with pain only as a presenting symptom of nerve injury may not have been identified. Women with stillborn fetuses were not included in this study, as most of them were discharged from the hospital before the initial interview. A physiatrist did not examine eight women. In addition, the diagnoses were not confirmed by quantitative studies (eg, electromyography), and therefore, the degree of axonal loss was not determined. On physical examination it may be difficult to differentiate between radicular, lumbosacral, and peripheral nerve lesions. However, as most of the injuries in the current study improved and resolved spontaneously, further study was not warranted for clinical reasons. Also, electromyography only measures large nerve fiber changes and may take as long as 3 weeks after injury to show changes. Some of the
OBSTETRICS & GYNECOLOGY
women’s symptoms would have resolved at the time of electromyography testing, even if it had been used. In addition, nerve injuries that were asymptomatic at the time of the initial interview (8 –32 hours after delivery) were not identified. Pushing position data were missing for some women. Directed interviews and examinations were not performed before delivery, and the presence of a preexisting injury was only documented by history after delivery. It is likely that different nerve injuries have different risk factors, and that analyzing all the injuries together may not have identified these factors. Our logistic regression model was a poor predictor of nerve injury, and there are likely other factors that we did not analyze that contribute to risk (eg, fetal head position or anatomic variation in nerve pathway or pelvic shape).8 Although the mechanism of nerve injury is not clear, we believe consideration should be given to changing the position of the lower extremities frequently during prolonged pushing, avoidance of prolonged thigh flexion, avoidance of extreme thigh abduction and external rotation, and shortened active pushing time by allowing the fetus to descend to the perineum without active maternal pushing.32
REFERENCES 1. Ong BY, Cohen MM, Esmail A, Cumming M, Kozody R, Palahniuk RJ. Paresthesias and motor dysfunction after labor and delivery. Anesth Analg 1987;66:18–22. 2. Vargo MM, Robinson LR, Nicholas JJ, Rulin MC. Postpartum femoral neuropathy: Relic of an earlier era? Arch Phys Med Rehabil 1990;71:591–6. 3. Scott DB, Tunstall ME. Serious complications associated with epidural/spinal blockade in obstetrics: A two-year prospective study. Int J Obstet Anesth 1995;4:133–9. 4. Holdcroft A, Gibberd FB, Hargrove RL, Hawkins DF, Dellaportas CI. Neurological complications associated with pregnancy. Br J Anaesth 1995;75:522–6. 5. Dar AQ, Robinson APC, Lyons G. Postpartum neurologic symptoms following regional blockade: A prospective study with case controls. Int J Obstet Anesth 2002;11: 85–90. 6. Institute of Medicine. Nutrition during pregnancy: I, weight gain; II, nutrient supplements. Washington: National Academy Press, 1990. 7. Operative vaginal delivery. Technical bulletin no. 196. In: 2000 compendium of selected publications. Washington: American College of Obstetricians and Gynecologists, 1994:644 –9. 8. Donaldson JO. Neurology of pregnancy. Philadelphia: WB Saunders, 1989.
VOL. 101, NO. 2, FEBRUARY 2003
9. Al-Hakim M, Katirji MB. Femoral mononeuropathy induced by the lithotomy position: A report of 5 cases with a review of the literature. Muscle Nerve 1993;16:891–5. 10. Rates of cesarean delivery—United States, 1993. MMWR Morb Mortal Wkly Rep 1995;44:303–7. 11. Hawkins JL, Beaty BR, Gibbs CP. Update on U.S. OB anesthesia practices. Anesthesiology 1999;91:A1060. 12. Zhang J, Klebanoff MA, DerSimonian R. Epidural analgesia in association with duration of labor and mode of delivery: A quantitative review. Am J Obstet Gynecol 1999;180:970–7. 13. Reynolds F. Maternal sequelae of childbirth. Br J Anaesth 1995;75:515–7. 14. Dawson DM, Krarup C. Perioperative nerve lesions. Arch Neurol 1989;46:1355–60. 15. Wilbourne AJ. Iatrogenic nerve injuries. Neurol Clin 1998;16:55–82. 16. Kuntzer T, VanMelle G, Regli F. Clinical and prognostic features in unilateral femoral neuropathies. Muscle Nerve 1997;20:205–11. 17. Haerer AF. DeJong’s the neurologic examination. Philadelphia: J. B. Lippincott, 1992. 18. Aminoff MJ. Neurological disorders and pregnancy. Am J Obstet Gynecol 1978;132:325–35. 19. Redick LF. Maternal perinatal nerve palsies. Postgrad Obstet Gynecol 1992;12:1–5. 20. Brasch RC, Bufo AJ, Kreienberg PF, Johnson GP. Femoral neuropathy secondary to the use of self-retaining retractor. Dis Colon Rectum 1995;38:1115–8. 21. Biemond A. Femoral neuropathy. In: Vinken PJ, Bruyn GW, eds. Volume 8: Handbook of clinical neurology. New York: John Wiley & Sons, 1977:303–10. 22. Roblee MA. Femoral neuropathy from the lithotomy position: Case report and new leg holder for prevention. Am J Obstet Gynecol 1967;97:871–2. 23. Spratt JD, Logan BM, Abrahams PH. Variant slips of psoas and iliacus muscles, with splitting of the femoral nerve. Clin Anat 1996;9:401–4. 24. Weinreb JC, Wolbarsht LB, Cohen JM, Brown CE, Maravilla KR. Prevalence of lumbosacral intervertebral disk abnormalities on MR images in pregnant and asymptomatic nonpregnant women. Radiology 1989;170:125–8. 25. Paech MJ, Godkin R, Webster GS. Complications of obstetric epidural analgesia and anaesthesia: A prospective analysis of 10,995 cases. Int J Obstet Anesth 1998;7:5–11. 26. Colachis SC 3rd, Pease WS, Johnson EW. A preventable cause of foot drop during childbirth. Am J Obstet Gynecol 1994;171:270–2. 27. Adornato BT, Carlini WG. “Pushing palsy:” A case of self-induced bilateral peroneal palsy during natural childbirth. Neurology 1992;42:936–7. 28. Reif ME. Bilateral common peroneal nerve palsy secondary to prolonged squatting in natural childbirth. Birth 1988;15:100–2.
Wong et al
Postpartum Nerve Injury
287
29. Babayev M, Bodack MP, Creatura C. Common peroneal neuropathy secondary to squatting during childbirth. Obstet Gynecol 1998;91:830–2. 30. Silva M, Mallinson C, Reynolds F. Sciatic nerve palsy following childbirth. Anaesthesia 1996;51:1144–8. 31. Warner MA, Warner DO, Harper M, Schroeder DR, Maxson PM. Lower extremity neuropathies associated with lithotomy positions. Anesthesiology 2000;93:938–42.
288
Wong et al
Postpartum Nerve Injury
32. Hansen SL, Clark SL, Foster JC. Active pushing versus passive fetal descent in the second stage of labor: A randomized controlled trial. Obstet Gynecol 2002;99:29–34. Reprints are not available. Received July 1, 2002. Received in revised form August 23, 2002. Accepted August 29, 2002.
OBSTETRICS & GYNECOLOGY
The reported incidence of postpartum lower extremity motor and sensory dysfunction secondary to neurological injury in present day obstetric practice is between 0.008% and 0.5%.1–5 The incidence has been inversely From the Departments of Anesthesiology and Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University, Chicago, Illinois. SD is currently affiliated with Rush Medical College, Chicago, Illinois, and HP with Washington University School of Medicine, St. Louis, Missouri.
related to sample size and varies widely with study methodology. In a retrospective review of 23,827 deliveries over a 9-year period the incidence of paresthesias and motor dysfunction was found to be 18.9 per 10,000 deliveries.1 A second retrospective review of 143,019 live births over a 16-year period reported an incidence of 0.8 per 10,000 births.2 Similarly, two prospective physician-reported audits from the early 1990s of 467,491 and 48,066 deliveries identified nerve injury rates of one per 10,000 and four per 10,000 respectively.3,4 In contrast, in a prospective, case-controlled study of 3341 parturients who received regional analgesia or anesthesia for labor and delivery, symptoms of nerve injury were reported at a rate of 58 per 10,000 deliveries.5 Anecdotal experience at our institution, based on neurological symptoms uncovered during routine postanesthetic visits, indicated that the incidence of nerve injury was higher than that estimated in previous studies. We hypothesized that the true incidence of postpartum nerve injury was higher than that reported in previous studies. The purpose of this study was to prospectively estimate the incidence of postpartum lumbosacral spine and lower extremity nerve injury in parturients receiving present day obstetric care, to assess the severity and duration of injury, and to determine factors associated with injury. MATERIALS AND METHODS The study was approved by the Institutional Review Board and spanned a 1-year period from July 1997 to June 1998. The calendar day after delivery (8 –32 hours after delivery), every parturient who delivered a liveborn infant at Prentice Women’s Hospital was asked a standard question by one of the investigators: “Do you have any leg numbness or weakness?” Spanish-speaking women were asked to read the question on a card. Interpreters were used for other languages. If the answer was affirmative or equivocal, the investigator performed a preliminary neurological examination to ascertain that neurological pathology was probably present. She then
VOL. 101, NO. 2, FEBRUARY 2003 © 2003 by The American College of Obstetricians and Gynecologists. Published by Elsevier.
0029-7844/03/$30.00 doi:10.1016/S0029-7844(02)02727-8
279
asked the woman for informed, written consent to participate in study follow-up, which consisted of a complete lower extremity neurological examination by a physiatrist investigator within 24 hours of interview and further diagnostic workup and referrals for physical therapy and rehabilitation, as clinically indicated. Thereafter, women with confirmed nerve injury (as diagnosed by physical examination performed by a physiatrist) were contacted by phone by a research nurse every week for 4 weeks, and then every month until symptom resolution. Women who answered in the affirmative regarding leg numbness or weakness but who refused follow-up were considered to have suspected but not confirmed nerve injury. The following information was documented on admission (by the nursing staff) or immediately postpartum (by the nursing staff, delivering obstetrician, or midwife) and was abstracted from an institutional database for analysis: maternal age, height, weight, prepregnancy weight and pregnancy weight gain, parity, type of labor analgesia, duration of second stage of labor, fetal presentation (vertex or nonvertex), type of delivery (spontaneous, vacuum- or forceps-assisted vaginal, or elective or scheduled versus nonscheduled cesarean delivery), and newborn weight. During the study period, the nursing staff documented fetal station and position at the time maternal pushing was initiated, and duration of pushing. Pushing position was documented in epochs of 15 minutes duration with the primary position during the epoch (lateral, supine, Fowler–lithotomy with thigh flexion less than 90° or with thigh flexion greater than 90°, pushing stool, pushing bar, squatting, or kneeling). Pregnancy weight gain was classified as low, normal, or high as defined by the National Academy of Science Institute of Medicine criteria.6 Prolonged second stage of labor was defined using The American College of Obstetricians and Gynecologist’s criteria, depending on parity and the presence or absence of neuraxial labor analgesia.7 Data were compared between women with elective or scheduled cesarean delivery and those with all other deliveries by 2, the two-sample t test, or the two-tailed Kolmogorov–Smirnov two-sample test. The proportion of women reporting symptoms of nerve injury were compared between women with elective or scheduled cesarean delivery and those with all other methods using a 2 statistic. In women who labored, univariable analysis of parturient and fetal characteristics believed to be associated with the potential for nerve injury were compared between women with confirmed nerve injury and those without by the 2 or Fisher exact test (parity [nulliparous versus parous], regional analgesia, fetal position [vertex versus nonvertex], prolonged second stage
280
Wong et al
Postpartum Nerve Injury
of labor, pregnancy weight gain [low, normal, high], prepregnancy body mass index [BMI] [underweight, normal, overweight6], infant weight [greater than 50th percentile], method of delivery, and gestational age [term versus preterm]) or the two-sample t test or the two-tailed Kolmogorov–Smirnov two-sample test (maternal age, height, weight, and BMI at time of delivery). A P value of less than .05 was required to reject the null hypothesis. Logistic regression analysis was performed using a backward-stepwise likelihood ratio elimination method with maternal and fetal variables for which P ⬍ .2 on univariable analysis. Models were developed for confirmed nerve injuries, and confirmed plus suspected nerve injuries. Criteria for entry and removal from the equation were set at .05 and .1, respectively. Concordance of the model with the observed outcomes was determined by measuring the area under the receiver operating characteristic curve. Bivariate correlation between duration of pushing and duration of the second stage of labor was determined using the Spearman . Box plots of the number of 15-minute epochs spent pushing in semi-Fowler–lithotomy less than 90°, greater than 90°, and other positions (identifying median and interquartile range, tenth to 90th percentile range, and mean) were constructed to compare women with nerve injury and those without nerve injury. The number of epochs in each pushing position in parturients with and without confirmed nerve injury was compared using the Mann–Whitney U test. Kaplan–Meier survival curves (percentage of women in each group with continuing symptoms of nerve injury) were constructed, and the median duration of symptoms was determined. The log rank test was used to compare symptom duration in the femoral nerve, lateral femoral cutaneous, and other nerve distributions, and to compare duration of symptoms involving sensory-only versus sensory–motor deficits. RESULTS Six thousand fifty-seven mothers delivered live-born infants during the study period. The demographic, labor, and fetal characteristics of these 6057 women are listed in Table 1. There were multiple differences between women who had an elective or scheduled cesarean delivery and those who did not. Nine women were not queried because of a language barrier or early discharge from the hospital, but were assumed not to have nerve injury for purposes of data analysis. The incidence of nerve injury is delineated in Table 2. Women with nerve injuries that were preexisting (by history) or not confirmed by physiatrist’s examination were considered without nerve injury. The incidence of
OBSTETRICS & GYNECOLOGY
Table 1. Parturient Characteristics
Maternal Age (y) Height (in.) Prepregnancy weight (kg) Prepregnancy BMI (kg · m⫺2) Weight gain (kg) Delivery weight (kg) Nulliparous (%) Gestational age (wk) Labor analgesia None (%) Systemic only (%) Regional only (%) Systemic and regional (%) Infant presentation* Cephalad OA (%) Cephalad OP/OT (%) Transverse lie (%) Breech (%) Other (%) Weight (g) Multiple gestation (%)
Labor (n ⫽ 5603)
No labor (n ⫽ 454)
P
30.2 ⫾ 5.9 65 ⫾ 3 63 ⫾ 13 23.5 ⫾ 4.8 14 ⫾ 5 78 ⫾ 14 51.7 39.0 ⫾ 2.2
32.5 ⫾ 5.7 65 ⫾ 3 67 ⫾ 18 25.0 ⫾ 6.6 14 ⫾ 5 81 ⫾ 17 43.0 37.8 ⫾ 2.8
.001 .710 .001 .001 .540 .001 .001 .001
13 15 55 17
NA NA NA NA
90.7 7.5 0.2 1.4 0.2 3347 ⫾ 598 2.3
58.3 0.9 3.8 36.2 0.9 3188 ⫾ 799 9.0
}
.001 .001 .001
BMI ⫽ body mass index; OA ⫽ occiput anterior; OP ⫽ occiput posterior; OT ⫽ occiput transverse. * Immediately before delivery.
confirmed nerve injury for women undergoing elective or scheduled cesarean delivery (0.22%) did not differ from that for all other women (0.98%) (P ⫽ .125, Fisher exact test). The relative risk of nerve injury for laboring women versus nonlaboring women in this study was 0.99 (95% confidence interval 0.99, 1.00). The distribution of nerve injuries is listed in Table 3. All diagnoses were made by physical examination. Lateral femoral cutaneous neuropathies, followed by femoral neuropathies, were the most common findings. Radiculopathies were at the L4, L5, and S1 dermatome levels. Overall, 21 women had a motor deficit, an incidence of 0.37%. Univariable analysis comparing laboring women with confirmed nerve injury to laboring women without nerve injury showed that laboring women with postparTable 2. Incidence of Nerve Injuries
Eligible participants (N ⫽ 6057) Interviewed Symptoms on interview Examined by physiatrist Findings [n (%)] Preexisting Not confirmed New confirmed
VOL. 101, NO. 2, FEBRUARY 2003
Labor [n (%)]
No labor [n (%)]
5603 (92.5) 5594 (99.9) 80 (1.4) 72 (90)
454 (7.5) 454 (100) 4 (0.9) 4 (100)
9 (12.5) 8 (11) 55 (76.5)
1 (25) 2 (50) 1 (25)
tum nerve injury were more likely to be nulliparous, have prolonged second stage of labor, and have assisted (forceps or vacuum) vaginal deliveries (Table 4). There were no differences between laboring women with nerve injury and those without injury in maternal age, height, weight, prepregnancy BMI, pregnancy weight gain, gestational age, fetal presentation, use of regional labor analgesia, or newborn weight. Backward stepwise logistic regression analysis identified nulliparity and prolonged second stage of labor as factors associated with confirmed nerve injury (Table 5). Concordance of the model as determined by the area under the receiver operating characteristic curve was 0.65. When the eight women with suspected nerve injury who were not examined by a physiatrist (because of discharge from the hospital before an examination [n ⫽ 7] or refusal to participate in the study [n ⫽ 1]) were analyzed with the confirmed injuries, regional labor analgesia was included in the model (Table 5). Concordance of this model with observed outcome was 0.67. Pushing data (pushing time and position, fetal station at initiation of pushing) were available for 70% of parturients without injury and 91% of parturients with nerve injury. The position of the fetal presenting part at initiation of pushing was not documented for the majority of parturients; therefore the available data were not analyzed. Little time was spent in pushing positions other
Wong et al
Postpartum Nerve Injury
281
Table 3. Distribution of Nerve Injuries in Laboring Parturients New Lateral femoral cutaneous Unilateral Bilateral Femoral (sensory deficit only) Unilateral Femoral (sensory and motor deficit) Unilateral Bilateral Common peroneal† Unilateral Lumbosacral plexus Unilateral Sciatic Unilateral Obturator Unilateral Radicular Unilateral Other Unilateral Totals
Total new
Confirmed
Suspected*
Preexisting
22 2
21 2
1 0
8 0
8
5
3
0
13 1
13 1
0 0
0 0
3
3
0
0
3
3
0
0
2
1
1
1
3
0
3
0
5
5
0
0
1 63
1 55
0 8
1 10
* Women reporting new symptoms of nerve injury who were not examined by a physiatrist. † Also known as the fibular nerve.
than semi-Fowler–lithotomy less than 90° thigh flexion and semi-Fowler–lithotomy greater than 90° thigh flexion; therefore these positions were combined into an “other” category for purposes of data analysis. There was a difference in the number of 15-minute epoch parturients with nerve injuries spent in either semi-Fowler–lithotomy position versus parturients with no nerve injury (Figure 1). The mean difference in pushing time and duration of second stage was 9 ⫾ 36 minutes, and pushing time was highly correlated with the duration of the second stage of labor ( ⫽ 0.912, P ⬍ .01). Women with nerve injuries began pushing with the fetal presenting part at a significantly higher station (P ⫽ .03) than women without nerve injury (Figure 1), and fetal station was correlated with both pushing time ( ⫽ ⫺0.3, P ⬍ .01) and duration of the second stage of labor ( ⫽ ⫺0.3, P ⬍ .001). The median duration of nerve injury symptoms was 2 months (range 1 week to greater than 18 months) (Figure 2). There was no difference (log rank 0.76, P ⫽ .68) in the median duration of symptoms when compared by nerve distribution (Figure 2, inset), or sensory-only injuries compared with sensory–motor (log rank 0.01, P ⫽ .93). Nine women were lost to follow-up (lateral femoral cutaneous nerve [three], femoral nerve [two], and other [four]). Two women continued to have symptoms (occasional pain) after 14 and 18 months, respectively. Six of the eight patients with suspected nerve injury received
282
Wong et al
Postpartum Nerve Injury
follow-up phone calls, with the duration of symptoms lasting between 2 weeks and 18 months (median 2 months). DISCUSSION Neurological injury associated with pregnancy and delivery, or intrinsic obstetric palsy, has long been recognized. The incidence of femoral neuropathy around 1900 was 3.2%, and 25% of cases were bilateral.8 Historically, case reports of postpartum nerve injury were associated with primiparity, cephalopelvic disproportion, and midforceps deliveries.8 The results of the current study support historical experience, in that nulliparas with prolonged second stage of labor were more likely to experience postpartum nerve injury, and that the majority of these nerve injuries were intrinsic obstetric palsies and were not directly related to regional anesthesia or analgesia.8 The incidence of postpartum injury in the current study was 1%, higher than the estimated incidence from retrospective studies performed between the late 1970s and early 1990s,1–3 but on the same order of magnitude as a recent prospective study in which postpartum women who received regional analgesia or anesthesia were asked a question similar to that in the current study (0.57%).5 There may be several reasons for the large difference in incidences found in previous studies per-
OBSTETRICS & GYNECOLOGY
Table 4. Univariable Analysis of Risk Factors in Laboring Parturients
Prolonged second stage (%) No Yes Nulliparity (%) No Yes Delivery (%) NSVD Vaginal, assisted Cesarean Regional labor analgesia (%) No Yes Newborn weight ⬎ 50th percentile (%) No Yes Prepregnancy BMI (kg 䡠 m⫺2) (%)† ⬍19.8 ⱖ19.8 and ⱕ26.0 ⬎26.0 Presentation/position (%) Cephalad, OA Cephalad, OP Other Gestation (%) ⬍37 wk ⱖ37 wk Weight gain (%)‡ Low Normal High Height (in.) Current weight (kg) Age (y)
No nerve injury (n ⫽ 5548)*
Confirmed nerve injury (n ⫽ 55)
90.9 8.1
76.4 23.2
.001
48.6 51.4
25.5 74.5
.001
81.7 9.6 8.7
69.1 20.0 10.9
28.0 72.0
16.4 83.6
.055
54.2 45.8
43.6 56.4
.116
18.4 60.0 21.6
27.8 48.1 24.1
.136
90.5 7.7 1.8
83.6 14.6 1.8
.163
8.4 91.6
3.6 96.4
23.7 38.6 37.7 64.6 ⫾ 2.8 77 ⫾ 14 30 ⫾ 6
23.4 48.9 27.7 64.3 ⫾ 2.7 79 ⫾ 18 30 ⫾ 6
P
.025
.205
.283 .375 .461 .778
NSVD ⫽ normal spontaneous vaginal delivery; vaginal assisted ⫽ forceps- or vacuum-assisted vaginal delivery. Other abbreviations as in Table 1. * Includes women with symptoms on interview not examined by physiatrist. † Underweight, normal, or overweight as defined by the Institute of Medicine.6 ‡ As defined by the Institute of Medicine.6
formed in the present era of obstetric care and that in the current study. The reported incidence of nerve injury 2 decades ago may have been falsely low.5,9 Retrospective
studies have known limitations. For example, none of the women diagnosed with nerve injury in our study were coded for such a diagnosis in their medical record.
Table 5. Multivariable Predictors of Nerve Injury in Laboring Parturients Variable Nulliparity Prolonged labor Constant Nulliparity Prolonged labor Regional labor analgesia Constant

SE
P
Exp ()
Women with confirmed nerve injury 0.94 0.34 .005 2.55 0.84 0.38 .025 2.32 ⫺5.41 0.29 .000 0.004 Women with suspected and confirmed nerve injury 0.95 0.30 .002 2.57 1.07 0.31 .001 2.92 0.69 0.36 .057 2.00 ⫺5.80 0.40 .000 0.003
95% CI for Exp () Lower
Upper
1.32 1.11
4.94 4.84
1.43 1.59 0.98
4.64 5.36 4.08
SE ⫽ standard error.
VOL. 101, NO. 2, FEBRUARY 2003
Wong et al
Postpartum Nerve Injury
283
Figure 1. Box plot of the number of 15-min epochs spent pushing in semi-Fowler–lithotomy less than 90°, greater than 90°, and other positions, and fetal station at initiation of pushing, for women with nerve injury versus those without nerve injury. Median shown as solid line and mean as dotted line, with box representing 25th to 75th percentile and whiskers tenth and 90th percentiles, respectively. Data are compared between groups using the Mann–Whitney U test. †P ⬍ .05. Wong. Postpartum Nerve Injury. Obstet Gynecol 2003.
Therefore, a retrospective search by diagnosis code2 would have yielded an incidence of zero. Vargo postulated that a change in obstetric management (increase in cesarean delivery rate, decrease in the use of midrotational forceps, and a decrease in the duration of the second stage of labor) contributed to a reduced incidence of nerve injury during the 1970s and 1980s relative to reports from earlier in the 20th century. The national cesarean delivery rate increased markedly during the 1970s and 1980s (17.5% in 1988); however, it declined during the 1990s,10 particularly at our institution, where the primary cesarean delivery rate at the time of the study was 10.8%. During this same period the national rate of regional labor analgesia in large hospitals increased from 16% in 1981 to 66% in 199711 and was 72% in our institution at the time of this study. Although regional labor analgesia was not directly associated with confirmed nerve injury in the current study, a large proportion of women received it, and therefore a larger study with more women not receiving regional analgesia would be necessary to adequately address whether regional labor analgesia is associated with postpartum nerve injury. Regional labor analgesia is associated with a longer
284
Wong et al
Postpartum Nerve Injury
second stage of labor,12 and parturients with labor analgesia–induced sensory blockade may not appreciate symptoms of impending nerve injury13 and may not change their body position in a timely manner. The presence of regional analgesia may encourage parturients to push without changing body position frequently, or may have contributed to an overall change in the proportion of time parturients push in various positions. Our data do not allow us to differentiate whether prolonged pushing (with flexion and external rotation of the thighs) or prolonged second stage is associated with nerve injury because the duration of pushing in this study was highly correlated with the duration of the second stage. Most parturients at our institution are instructed to push once the cervix is completely dilated. Acute injury to nerves can occur by several mechanisms, including transection, stretch, compression, or vascular injury.14 Compression and stretch may result in decreased perineural blood flow and ischemia. Less severe injuries cause focal demyelination and conduction block, whereas more severe injuries cause axon loss resulting in nerve conduction failure.15 Recovery of nerve function is best predicted by the percentage of axonal loss, with axonal loss less than 50% predicting
OBSTETRICS & GYNECOLOGY
Figure 2. Kaplan–Meier plot of the duration of neurological symptoms in women with confirmed new nerve injuries. Inset panel shows duration of symptoms by nerve distribution. Wong. Postpartum Nerve Injury. Obstet Gynecol 2003.
recovery within 1 year.16 Based on the duration of their symptoms, the women in this study likely suffered minor degrees of axon loss or focal demyelination. Interestingly, Ong and colleagues1 found that symptoms of nerve injury had resolved within 72 hours in the 45 women they identified, compared with a median duration of symptoms of 2 months in the current study. Others have found that symptoms usually resolve in less than 6 months.5 The majority of women had injury to the lateral femoral cutaneous nerve, also known as meralgia paresthetica. Risk factors for lateral femoral cutaneous neuropathy include anatomic variation (the nerve may bisect the inguinal ligament), increased intraabdominal pressure, pregnancy, obesity, diabetes mellitus, trauma, pressure by a belt, or prolonged hip flexion.17 Exacerbated lumbar lordosis (eg, during pregnancy) leads to compression of the nerve by the posterior fascicle of the inguinal ligament in patients in whom the nerve bisects the inguinal ligament.18 Indeed, nine parturients identified in this study described symptoms consistent with this syndrome before delivery. The nerve may be entrapped or compressed by the inguinal ligament during maternal pushing in the thigh-flexed position. Interestingly, four women with this neuropathy underwent a
VOL. 101, NO. 2, FEBRUARY 2003
cesarean delivery before the second stage of labor. The nerve may have been injured during cesarean delivery, by cutting (a wide Pfannenstiel incision), stretching, or pressure by a retractor.19 Alternatively, laboring parturients usually wear one to two elastic belts around their lower abdomen to hold monitors in place. Perhaps these belts or monitors applied pressure on the nerve. The second most common injury was to the femoral nerve. Injury to the femoral nerve has been described after lower abdominal surgery using a Pfannenstiel incision, particularly with self-retaining retractors.20 However, no women with femoral nerve injury in our study had a cesarean delivery. The femoral nerve does not descend into the true pelvis, and therefore it is unlikely that compression by the fetal head or body is a mechanism of injury. The nerve may be compressed at the inguinal ligament during thigh flexion, external rotation, and abduction. However, if weakness of thigh flexion is present (iliopsoas muscles), then the lesion must be proximal to the inguinal ligament, as nerve fibers to the iliopsoas muscle branch proximal to the inguinal ligament.21 All 13 women in the current study with a motor component to their femoral neuropathy had weakness of hip flexion as well as knee extension, thereby placing the likely site of injury proximal to the inguinal ligament.
Wong et al
Postpartum Nerve Injury
285
Femoral nerve injury with iliopsoas weakness has been described after vaginal hysterectomies in the lithotomy position, and cadaver studies showed that retractor replacement was not responsible for nerve injury.21 The intrapelvic portion of the femoral nerve has a poor blood supply relative to other portions of the nerve.21 Several authors have suggested that excessive hip abduction and external rotation may cause a stretch lesion of the intrapelvic portion of the nerve22 and that poor blood supply may contribute to stretch-induced nerve ischemia.9 Finally, variant slips of the psoas and iliacus muscles that split the femoral nerve have been identified. They may place tension on the femoral nerve and lead to an entrapment syndrome.23 Five laboring women presented with history and physical examination findings most consistent with injuries at the radicular level. Radicular symptoms can be caused by many processes, including bulging or ruptured intervertebral disks and, less commonly, arachnoiditis, spinal cord, intra- or extradural tumors, trauma, infection, inflammation, or even muscle spasm.17 There is contradictory evidence as to whether pregnancy and labor predispose to intervertebral disk prolapse and subsequent radiculopathy.24 Although all five women in the current study had regional labor analgesia (three had combined spinal– epidural analgesia, and two had epidural analgesia), no women had paresthesias or pain during epidural or spinal needle insertion or drug injection. In contrast to the laboring patients diagnosed with radiculopathy, the single woman who had a nerve injury after elective cesarean delivery had postpartum radicular pain in the same distribution as a paresthesia elicited during initiation of spinal anesthesia. Five days after delivery her symptoms were improving, but she was lost to further follow-up. Recent prospective studies in obstetric patients have found a similarly low incidence of traumatic neuropathy (one in 10,99525 and one in 13,0074). In the middle of the 20th century foot drop was the most common obstetric nerve palsy.8 Foot drop can be secondary to injury at the level of the lumbosacral root, lumbosacral plexus, sciatic nerve, or the common peroneal nerve (also known as the fibular nerve). Even if the injury occurs at the level of the lumbosacral plexus, the common peroneal nerve component is more often affected than the tibial component of the plexus because of its relationship to the bony pelvis.18 Components of the lumbosacral plexus cross the pelvic brim or originate in the true pelvis and could be compressed between the bony pelvis and fetal head (or forceps). Indeed, fetal macrosomia; malpresentations such as occiput posterior, or brow; or specific pelvic features such as a straight
286
Wong et al
Postpartum Nerve Injury
sacrum or wide, posterior pelvis are described as risk factors for obstetric-related foot drop.18 The most common mechanism of peroneal nerve injury is external compression15 classically from pressure exerted by inappropriate leg positioning in stirrups. Two case reports describe postpartum peroneal neuropathy in women with prolonged pushing with the knees hyperflexed and the parturients’ hands grabbing and applying pressure over the lateral, upper leg.26,27 The nerve may also be injured during prolonged squatting during labor.28,29 Injury to the obturator nerve has also been described in the postpartum period. The nerve crosses the pelvic brim and may be compressed between the pelvis and fetal head or forceps applied to the fetal head.8,18 The lithotomy position worsens angulation of the nerve as it leaves the obturator foramen.18 Pudendal nerve blocks have caused hematomas that have led to entrapment neuropathies of the obturator nerve.8 Similar to this study, postpartum injury to the sciatic nerve has been reported,30 although the mechanism remains unclear.18,30 Stretch injuries to the sciatic nerve after gynecologic procedures in the lithotomy position have been described.18 In a prospective study of lower extremity neuropathies associated with 991 surgical procedures performed in the lithotomy position,31 the incidence of nerve injury was 1.5% and included obturator, lateral femoral cutaneous, sciatic, and peroneal nerve injuries. In agreement with our findings, nerve injury was associated with time in the lithotomy position, but was not associated with patient body habitus or diabetes mellitus. Symptoms resolved in 93% of patients by 6 months. There are several limitations to our study. The current study may have underestimated the incidence of lower extremity nerve injuries for a number of reasons. Some women may not have reported symptoms of nerve injury when interviewed, and women with pain only as a presenting symptom of nerve injury may not have been identified. Women with stillborn fetuses were not included in this study, as most of them were discharged from the hospital before the initial interview. A physiatrist did not examine eight women. In addition, the diagnoses were not confirmed by quantitative studies (eg, electromyography), and therefore, the degree of axonal loss was not determined. On physical examination it may be difficult to differentiate between radicular, lumbosacral, and peripheral nerve lesions. However, as most of the injuries in the current study improved and resolved spontaneously, further study was not warranted for clinical reasons. Also, electromyography only measures large nerve fiber changes and may take as long as 3 weeks after injury to show changes. Some of the
OBSTETRICS & GYNECOLOGY
women’s symptoms would have resolved at the time of electromyography testing, even if it had been used. In addition, nerve injuries that were asymptomatic at the time of the initial interview (8 –32 hours after delivery) were not identified. Pushing position data were missing for some women. Directed interviews and examinations were not performed before delivery, and the presence of a preexisting injury was only documented by history after delivery. It is likely that different nerve injuries have different risk factors, and that analyzing all the injuries together may not have identified these factors. Our logistic regression model was a poor predictor of nerve injury, and there are likely other factors that we did not analyze that contribute to risk (eg, fetal head position or anatomic variation in nerve pathway or pelvic shape).8 Although the mechanism of nerve injury is not clear, we believe consideration should be given to changing the position of the lower extremities frequently during prolonged pushing, avoidance of prolonged thigh flexion, avoidance of extreme thigh abduction and external rotation, and shortened active pushing time by allowing the fetus to descend to the perineum without active maternal pushing.32
REFERENCES 1. Ong BY, Cohen MM, Esmail A, Cumming M, Kozody R, Palahniuk RJ. Paresthesias and motor dysfunction after labor and delivery. Anesth Analg 1987;66:18–22. 2. Vargo MM, Robinson LR, Nicholas JJ, Rulin MC. Postpartum femoral neuropathy: Relic of an earlier era? Arch Phys Med Rehabil 1990;71:591–6. 3. Scott DB, Tunstall ME. Serious complications associated with epidural/spinal blockade in obstetrics: A two-year prospective study. Int J Obstet Anesth 1995;4:133–9. 4. Holdcroft A, Gibberd FB, Hargrove RL, Hawkins DF, Dellaportas CI. Neurological complications associated with pregnancy. Br J Anaesth 1995;75:522–6. 5. Dar AQ, Robinson APC, Lyons G. Postpartum neurologic symptoms following regional blockade: A prospective study with case controls. Int J Obstet Anesth 2002;11: 85–90. 6. Institute of Medicine. Nutrition during pregnancy: I, weight gain; II, nutrient supplements. Washington: National Academy Press, 1990. 7. Operative vaginal delivery. Technical bulletin no. 196. In: 2000 compendium of selected publications. Washington: American College of Obstetricians and Gynecologists, 1994:644 –9. 8. Donaldson JO. Neurology of pregnancy. Philadelphia: WB Saunders, 1989.
VOL. 101, NO. 2, FEBRUARY 2003
9. Al-Hakim M, Katirji MB. Femoral mononeuropathy induced by the lithotomy position: A report of 5 cases with a review of the literature. Muscle Nerve 1993;16:891–5. 10. Rates of cesarean delivery—United States, 1993. MMWR Morb Mortal Wkly Rep 1995;44:303–7. 11. Hawkins JL, Beaty BR, Gibbs CP. Update on U.S. OB anesthesia practices. Anesthesiology 1999;91:A1060. 12. Zhang J, Klebanoff MA, DerSimonian R. Epidural analgesia in association with duration of labor and mode of delivery: A quantitative review. Am J Obstet Gynecol 1999;180:970–7. 13. Reynolds F. Maternal sequelae of childbirth. Br J Anaesth 1995;75:515–7. 14. Dawson DM, Krarup C. Perioperative nerve lesions. Arch Neurol 1989;46:1355–60. 15. Wilbourne AJ. Iatrogenic nerve injuries. Neurol Clin 1998;16:55–82. 16. Kuntzer T, VanMelle G, Regli F. Clinical and prognostic features in unilateral femoral neuropathies. Muscle Nerve 1997;20:205–11. 17. Haerer AF. DeJong’s the neurologic examination. Philadelphia: J. B. Lippincott, 1992. 18. Aminoff MJ. Neurological disorders and pregnancy. Am J Obstet Gynecol 1978;132:325–35. 19. Redick LF. Maternal perinatal nerve palsies. Postgrad Obstet Gynecol 1992;12:1–5. 20. Brasch RC, Bufo AJ, Kreienberg PF, Johnson GP. Femoral neuropathy secondary to the use of self-retaining retractor. Dis Colon Rectum 1995;38:1115–8. 21. Biemond A. Femoral neuropathy. In: Vinken PJ, Bruyn GW, eds. Volume 8: Handbook of clinical neurology. New York: John Wiley & Sons, 1977:303–10. 22. Roblee MA. Femoral neuropathy from the lithotomy position: Case report and new leg holder for prevention. Am J Obstet Gynecol 1967;97:871–2. 23. Spratt JD, Logan BM, Abrahams PH. Variant slips of psoas and iliacus muscles, with splitting of the femoral nerve. Clin Anat 1996;9:401–4. 24. Weinreb JC, Wolbarsht LB, Cohen JM, Brown CE, Maravilla KR. Prevalence of lumbosacral intervertebral disk abnormalities on MR images in pregnant and asymptomatic nonpregnant women. Radiology 1989;170:125–8. 25. Paech MJ, Godkin R, Webster GS. Complications of obstetric epidural analgesia and anaesthesia: A prospective analysis of 10,995 cases. Int J Obstet Anesth 1998;7:5–11. 26. Colachis SC 3rd, Pease WS, Johnson EW. A preventable cause of foot drop during childbirth. Am J Obstet Gynecol 1994;171:270–2. 27. Adornato BT, Carlini WG. “Pushing palsy:” A case of self-induced bilateral peroneal palsy during natural childbirth. Neurology 1992;42:936–7. 28. Reif ME. Bilateral common peroneal nerve palsy secondary to prolonged squatting in natural childbirth. Birth 1988;15:100–2.
Wong et al
Postpartum Nerve Injury
287
29. Babayev M, Bodack MP, Creatura C. Common peroneal neuropathy secondary to squatting during childbirth. Obstet Gynecol 1998;91:830–2. 30. Silva M, Mallinson C, Reynolds F. Sciatic nerve palsy following childbirth. Anaesthesia 1996;51:1144–8. 31. Warner MA, Warner DO, Harper M, Schroeder DR, Maxson PM. Lower extremity neuropathies associated with lithotomy positions. Anesthesiology 2000;93:938–42.
288
Wong et al
Postpartum Nerve Injury
32. Hansen SL, Clark SL, Foster JC. Active pushing versus passive fetal descent in the second stage of labor: A randomized controlled trial. Obstet Gynecol 2002;99:29–34. Reprints are not available. Received July 1, 2002. Received in revised form August 23, 2002. Accepted August 29, 2002.
OBSTETRICS & GYNECOLOGY