ISCHIOPAGUS tetrapus conjoining (twins with a common pelvis, four legs, and shared urinary, genital, and intestinal organs) occurs in approximately 1 in 800,000 live births. 1Successful separation is still unusual, and survivors have significant long-term health problems. Pregnancy, although not recommended, can occur in some female survivors. We describe successful perioperative anesthetic management of cesarean delivery in a patient who is a separated ischiopagus twin.

A 23-yr old primigravid, separated ischiopagus tetrapus twin with hypoplastic mandible, uterine didelphys, and absent pelvic musculature presented in preterm labor at 35 weeks gestation. The patient had been joined to her sister at the pelvic rim. The twins had duplex urogenital sinus openings and a shared colon. The two girls were separated at 15 months of age at our institution. 2The twin died shortly thereafter because of recurrent episodes of pneumonia. Our patient underwent a total of 23 operations during general anesthesia between birth and her early teens. There were no anesthetic complications or airway difficulties during these operations. Surgeries included ileostomy, colostomy, formation of a nonneurogenic urinary neobladder, vaginoplasty, multiple perineoplasties with closure of vesicovaginal fistulae, repair of abdominal wall defects, leg lengthening, pelvic and hip reconstruction, and lower back surgery. The patient led an active life; she won a teen beauty contest, finished high school, and entered the work force as a supermarket cashier.

Her prenatal course was complicated by hyperemesis gravidarum and multiple urinary tract infections secondary to her altered urinary anatomy. She received nitrofurantoin for bacterial suppression for the final 10 weeks of her pregnancy.

Maternal weight gain added to the gait disturbances caused by the patient’s chronically enlarged intertrochanteric distance (1.5 times normal). This prevented her from pursuing her job near the end of pregnancy. Her reconstructed pelvic bones created an obstructed pelvic inlet that was best described as a broad flat oval. Vaginal delivery was not considered possible.

A magnetic resonance image of the pelvis without gadolinium contrast at 30 weeks gestation was ordered to describe the pelvic anatomy and to facilitate surgical planning. It revealed uterine didelphys. The pregnancy was located in the superior uterus, which abutted the abdominal wall. The placenta was posteriorly placed. The fetus had a vertex presentation. No fetal anomalies were noted. Inferior and anterior to the gravid uterus was a smaller, empty uterus; both uteri joined into a single vagina. The empty uterus was that of her twin, who after separation did not have a vagina or a uterus.

The patient was counseled to avoid pregnancy in the future. An elective cesarean delivery at 38 weeks in combination with tubal ligation was planned. However, the onset of preterm labor at 35 weeks hastened the delivery plan. Amniocentesis revealed unfavorable fetal lung indices, so dexamethasone was given to enhance fetal lung maturity. Cesarean delivery was performed 3 days later because the fetal heart rate pattern showed mild repetitive decelerations with regular uterine contractions.

There were no cardiovascular, respiratory, or neurologic symptoms. Physical examination revealed a 1.5-m-tall, 59-kg parturient. She had a hypoplastic mandible, a Mallampati score of 2, and a free range of movement in the temporomandibular joints and cervical spine. An ileostomy and multiple lower abdominal scars were noted. A vertical midline lumbosacral scar was present. There was no lumbar lordosis. Neurologic examination was normal. Full blood count, electrolytes, chest x-ray, electrocardiogram, and echocardiogram were within normal limits.

Little information was available about the patient’s previous back surgery; therefore, she underwent lumbosacral magnetic resonance imaging without gadolinium contrast to evaluate the feasibility of neuraxial anesthesia (figs. 1 and 2). The lumbar spine had normal alignment and segmentation with preserved disc spaces. The spinal cord seemed to be normal, with the conus medullaris terminating at T12–L1. A near-midline skin tag overlaid the L2–L3 interspace. At the L2–L3 interspace, the distances from the skin to the epidural space and the spinal nerves were 3.4 and 3.8 cm, respectively. Caudal images demonstrated not only a malformed pelvis with splayed iliac bones but also a midline dorsal defect in the caudal sacrum through which fat invaginated. No neural tissue was present in the sacrum, and the overlying soft tissues appeared normal.

Fig. 1. T2-weighted sagittal magnetic resonance image of lumbar spine at 35 weeks’ gestation. Annotations 3  and 4  at the L2–L3 level mark the distances from skin to epidural space (3.4 cm) and epidural to spinal nerves (0.4 cm), respectively.

Fig. 1. T2-weighted sagittal magnetic resonance image of lumbar spine at 35 weeks’ gestation. Annotations 3  and 4  at the L2–L3 level mark the distances from skin to epidural space (3.4 cm) and epidural to spinal nerves (0.4 cm), respectively.

Fig. 2. T2-weighted coronal magnetic resonance image of lumbar spine. Note the open sacral hiatus and partial sacral agenesis.

Fig. 2. T2-weighted coronal magnetic resonance image of lumbar spine. Note the open sacral hiatus and partial sacral agenesis.

The patient wanted to be awake during the delivery and requested that her surgery be performed during regional anesthesia. Because physical and radiologic evaluation suggested that neuraxial anesthesia was possible, we decided to attempt combined spinal–epidural anesthesia. We planned to change to a general anesthetic if the epidural or intrathecal spaces could not be identified, if the block level was too high or too low, or if the fetal status had suddenly deteriorated.

The skin tag noted on the magnetic resonance image allowed us to identify the L2–L3 interspace. In this patient, the Tuffier line would also have been an accurate landmark because it projected over the L4 spinous process. The patient sat during block placement. The epidural space was reached in a single attempt with an 18-gauge Tuohy needle at a depth of 4.0 cm. A 27-gauge Whitacre spinal needle was advanced through the epidural needle into the spinal fluid. Twelve milligrams hyperbaric spinal bupivacaine, 0.75%, and 100 μg morphine were injected. A 20-gauge catheter was inserted 5 cm into the epidural space. We laid the patient supine with left lateral tilt to displace the uterus off the inferior vena cava, which was in the normal position to the right of the vertebral column. A C6 sensory level was obtained within 12 min. The patient maintained good upper extremity strength throughout the case.

Surgery proceeded with a paramedian incision through the atrophic abdominal wall. The patient’s ileostomy functioned well throughout her pregnancy course and at the time of cesarean delivery was covered with an adhesive polyurethane membrane, which allowed direct visualization throughout the case. She had no rectus abdominus muscles and limited abdominal internal-oblique musculature. The abdominal musculature was thought to have been lost at the time of initial twin separation or was congenitally absent. Great care was taken at surgery to avoid these reconstructed areas, and additional efforts at abdominal closure were utilized to reinforce her attenuated fascial tissues with an overlapping technique to avoid future hernia formation.

A classic uterine incision was performed for visualization and to avoid the neobladder. A healthy male newborn (1- and 5-min Apgar scores of 9; weight, 2,200 g) was delivered. The solitary fallopian tube was ligated. Adhesions between the small bowel and the uterus were carefully lysed. The surgery lasted 66 min. Epidural dosing was not required, and the epidural catheter was removed at the end of surgery. Motor and sensory function returned to normal within 3 h.

Satisfactory postoperative analgesia was achieved with nonsteroidal analgesics in addition to the single dose of intrathecal morphine. The postpartum course was uneventful apart from a brief period of nausea. The mother was discharged from the hospital in good condition on postpartum day 3. Feeding difficulties and mild respiratory issues secondary to prematurity complicated the infant’s neonatal course. He was discharged home on day 9 from the neonatal intensive care unit, with a weight 2.18 kg.

The incidence of conjoined twins is 1 in 50,000 to 1 in 100,000 deliveries in the United States. 3Seventy percent of all conjoined twins are female. The ischiopagus type is rare, constituting only 6–11% of all conjoined twins. Only 63% of ischiopagus twins are separable. 4Each case involves sharing of various normal or anomalous parts of the urinary, genital, and lower digestive systems. Separate, complete vertebral columns are found in ischiopagus twins; however, there is a high incidence of hyper-extension, lordosis, scoliosis, and sacral fusion. In the tetrapus subtype of ischiopagus twinning, there are four legs and one large pelvic ring. Congenital vertebral anomalies (including thoracic hemivertebrae, missing lumbar vertebrae, and sacral agenesis) and visceral anomalies (including dextrocardia and lung hypoplasia) may accompany the pelvic and lower extremity problems. 1,5 

The only previously published record of a successful pregnancy in a separated conjoined twin is of a patient of the omphalopelvoischiopagus type. 5Cesarean delivery was performed to avoid disruption of the fragile pelvic floor and reconstructed perianal area. An abdominal x-ray film showed lumbar lordosis, a short sacrum, and iliac crests with old osteotomies and absent pubic rami. Surgery with a midline vertical skin incision and vertical fundal uterine incision proceeded during epidural anesthesia.

Magnetic resonance imaging delineated our patient’s nearly normal lumbar vertebral anatomy and gave us accurate information about the distance from skin to the epidural space. We choose to avoid using gadolinium contrast. Magnetopharmaceuticals such as gadolinium are considered safe for use during pregnancy, but these drugs are not innocuous. 6Clinical trials in pregnancy are limited by small numbers and lack of reliable data on neonatal and maternal outcome. 7–9The uncertainty of these risks outweighed in our minds the benefit of possible additional diagnostic information we could have obtained with gadolinium. The defect in closure of the sacral segments is most likely a result of the surgical separation of the bony fusion to the patient’s twin rather than representing a true, congenital spina bifida. Although congenital lumbosacral spinal anomalies are generally considered to be a contraindication to neuraxial anesthesia, epidural catheters have been safely placed in some spina bifida patients after radiologic imaging to outline spinal anatomy. 10 

As conjoined twin separation techniques improve, more patients will survive to develop adult medical problems. This is the first report of combined spinal–epidural anesthesia for cesarean delivery in a pregnant separated ischiopagus twin. Magnetic resonance imaging defined the pelvic and vertebral anatomy and assisted the planning of the surgical and anesthetic procedures. A successful outcome was achieved despite the patient’s multiple uterine, pelvic, and orthopedic anomalies.

Spiegel DA, Ganley TJ, Akbarnia H, Drummond DS: Congenital vertebral anomalies in ischiopagus and pygopagus conjoined twins. Clin Orthop 2000; 381:137–44
Shapiro E, Fair WR, Ternberg JL, Siegel MJ, Bell MJ, Manley CB: Ischiopagus tetrapus twins: Urological aspects of separation and 10-year followup. J Urol 1991; 145:120–5
O’Neill JA Jr, Holcomb GW III, Schnaufer L, Templeton JM Jr., Bishop HC, Ross AJ III, Duckett JW, Norwood WI, Ziegler MM, Koop CE: Surgical experience with thirteen conjoined twins. Ann Surg 1988; 208:299–312
Spencer R: Theoretical and analytical embryology of conjoined twins: II. Adjustments to union. Clin Anat 2000; 13:97–120
Shah LP, Chazotte C: Successful pregnancy in a separated conjoined twin. Am J Obstet Gynecol 1994; 171:1391–2
Carr JJ: Magnetic resonance contrast agents for neuroimaging: Safety issues. Neuroimaging Clin N Am 1994; 4:43–54
Lam G, Kuller J, McMahon M: Use of magnetic resonance imaging and ultrasound in the antenatal diagnosis of placenta accreta. J Soc Gynecol Invest 2002; 9:37–40
Spencer JA, Tomlinson AJ, Weston MJ, Lloyd SN: Early report: comparison of breath-hold MR excretory urography, Doppler ultrasound and isotope renography in evaluation of symptomatic hydronephrosis in pregnancy. Clin Radiol 2000; 55:446–53
Marcos HB, Semelka RC, Worawattanakul S: Normal placenta: Gadolinium-enhanced dynamic MR imaging. Radiology 1997; 205:493–6
Cooper MG, Sethna NF: Epidural analgesia in patients with congenital lumbosacral spinal anomalies. Anesthesiology 1991; 75:370–4