Surgical Problems in Pregnancy

The prevalence of major nongynecologic surgery or serious trauma during pregnancy is 1 to 10 per 1000 pregnancies. Minor trauma or domestic violence occurs in 10% to 30% of pregnancies. Gynecologic surgery, such as diagnostic laparoscopy (to rule out ectopic pregnancy), ovarian surgery, or cervical procedures (e.g., loop electrosurgical excision, cone biopsy) occur in an additional 1% to 3% of pregnancies (Table 1). This chapter primarily deals with nonobstetric surgical problems in pregnancy; however, it includes obstetric surgical conditions that could be a source of confusion in diagnosis or management.

Table 1. Estimated Frequency of Surgical Disease in Pregnancy per 1000 Births

The approach to surgery is different in pregnancy than in the nonpregnant state. A central focus is balancing the health and well-being of the fetus against the mother's need for surgery. The effect of surgery on the fetus and the pregnancy in general often is difficult to discern from the effects caused by the pathologic process (e.g., appendicitis) that created the need for surgery. However, there is a demonstrable increase in risk from surgery alone, and that risk to the fetus is greatest in the first and third trimester. There are conflicting data concerning the effect of first-trimester anesthesia on the rate of spontaneous abortion. There is a consistent, small increase in the likelihood of spontaneous abortion among women working in the operating room.¹ Mazze and associates² found no increase in spontaneous abortion among 3000 women who underwent first-trimester surgery. Duncan and colleagues,³ however, did report an increase. They stratified outcomes of 2565 women who underwent anesthesia during pregnancy by type of anesthesia and surgery. They reported an increased risk of spontaneous abortion among patients undergoing general anesthesia (risk ratio = 1.58) and obstetric and gynecologic surgery in particular (risk ratio = 2.0).

The major problem with retrospective studies on spontaneous abortion and general anesthesia is selection bias among cases and controls (selected for baseline rate of spontaneous abortion). The pathologic process that warranted surgery may have increased the patient's risk of spontaneous abortion. The timing and sensitivity of the pregnancy test may bias the selection of control patients. The risk of spontaneous abortion varies considerably by gestational age: the earlier the patient is tested, the higher the risk of spontaneous abortion. Most studies do not adequately match patients by gestational age. The sensitivity of pregnancy tests varies considerably (e.g., serum human chorionic gonadotropin versus latex agglutination tests—home pregnancy tests). The period of greatest discordance between serum and urine pregnancy tests is 3 to 5 weeks after the first day of the last menstrual period, and this also is the period of greatest risk of spontaneous abortion after a positive test result. Most studies have not evaluated timing and technique of pregnancy diagnosis.

The lessons of the thalidomide and diethylstilbestrol tragedies raise concerns about teratogenic effects of anesthetic agents. It is known that risk may be species or agent specific (e.g., humans and thalidomide) and that effects may manifest years after exposure (e.g., genital tract abnormalities and diethylstilbestrol). The risks to the fetus caused by maternal exposure to drugs are classified as follows: teratogenic (e.g., thalidomide), carcinogenic (e.g., diethylstilbestrol), growth retarding (e.g., cancer chemotherapy), and physiologic (e.g., β-sympathomimetic). The teratogenic risk of drugs is limited by gestational age. Toxic agents can cause spontaneous abortion or have no effect before 17 days after conception. Between 17 and 56 days after conception, the fetal organ systems are most susceptible to developmental injury (e.g., neural tube defects, congenital heart disease). Even if exposure to a teratogen occurs during the critical period, not all fetuses (usually less than 50%) are affected. The likelihood that a drug will cause developmental defects is related to genotype (family history of defects); the presence of other teratogens (e.g., smoking, alcohol, “street” drugs, medications); placental transfer; and the type, dose, and duration of exposure. After 56 postconception days, the major impact of exposure is the reduction of fetal cell number (i.e., symmetric growth retardation) or impaired fetal physiology.

Many different anesthetic agents are used alone and in combination. Although many of the most common agents (e.g., N₂O, diazepam [Valium], isoflurane) demonstrate species-specific teratogenic risk,² these results neither have been consistently reproduced nor easily translated to the human experience. Among 8000 women who underwent surgery during pregnancy, there was no demonstrated increased incidence of congenital abnormalities in the offspring.^3,4 To an extent, the controversy concerning first-trimester exposure to anesthesia is moot. The incidence of spontaneous abortion (15% to 20%) and congenital abnormalities (3% to 5% by 5 years of age) is high enough, and the propensity of patients and physicians to attach blame is strong enough, that elective surgery in the first 13 weeks after the first day of the last menstrual period is fraught with potential legal risk and psychological concerns.

After 13 weeks, the major organ systems of the fetus are developed; the risk of congenital malformation is minimal. Between 13 and 23 weeks, the uterus is less sensitive to the stimulating effects of surgery, and the risk of preterm labor is minimal. In addition, before 24 weeks heroic measures (i.e., cesarean section) would not be performed if preterm labor is caused by surgery. Therefore, if elective surgery must be performed during pregnancy, the period between 13 and 23 weeks is optimal.

After 24 weeks' gestation, surgery can produce three complications: fetal hypoxia, infection, and preterm labor and birth. The most common cause for fetal hypoxia is supine hypotension. After 16 weeks' gestation, when the uterus becomes an extra pelvic organ, the enlarging uterus has an increasing potential to obstruct return venous flow in the inferior vena cava when the gravida is supine. In the late third trimester, symptomatic hypotension (characterized by syncope, nausea, and vomiting) occurs in 10% of healthy gravidas. This risk is higher when the patient's cardiovascular system is further challenged by sepsis (e.g., appendicitis), hypovolemia (e.g., trauma), or sympathetic blockage (e.g., epidural analgesia). After 16 weeks, all patients undergoing surgery are placed in the left lateral tilt position to reduce venous and arterial compression by the pregnant uterus.

Fetal death and brain injury can occur at any gestational age and usually result from fetal hypoxia. Although prolonged, severe supine hypotension is one cause, many other causes are possible. Specific maternal findings measure the degree of risk. In particular, a decrease in maternal hematocrit of over 50% and a decrease in maternal mean blood pressure of 20%, or a maternal Pao₂ below 60 mmHg (O₂ saturation less than 90%),⁵ will result in fetal hypoxia, acidosis, and compromise. In addition to brain injury, the hypoxic stress to the fetus may stimulate the fetus to produce the hormonal triggers of preterm labor.

A significant proportion of fetal neurodevelopment occurs in the third trimester. Insult through the primary disease process, surgical complications, anesthetic agents, or anesthetic management (e.g., respiratory support) has the potential to affect neonatal and childhood neurodevelopment. Animal models have been used to investigate whether anesthetic agents can cause neurodevelopmental handicap. Exposure to local anesthetic or inhalation anesthetics has been associated with neurodevelopmental deficits in rodents.^1,2 Transfer to the human experience is fraught with error. The few studies that have been performed on children exposed to general anesthesia in utero yield conflicting data. A provocative study by Hollenbeck and associates⁶ observes that scores on one of three standardized intelligence tests were lower among 4-year-old children who were exposed to anesthetics in utero compared with the scores of unexposed children of the same age (91 ± 15 versus 108 ± 20, respectively [mean ± SD]). Human studies will never be able to rule out the possibility that the pathologic conditions requiring surgery may be the culprit in these neurologic deficits rather than the anesthetic agent or management.

Infection that results from the primary disease process (i.e., appendicitis) or as the result of surgery (i.e., wound infection) injures the fetus by direct extension or hematogenous inoculation (chorioamnionitis) or indirectly through preterm birth. Decidual macrophages initiate preterm labor by increased prostaglandin production. Circulating endotoxins or cytokines (tumor necrosis factor, interleukin-1, interleukin-6, platelet-activating factor) from distal infections stimulate the decidual macrophages to produce prostaglandins.⁷ Endotoxins and cytokines that transverse the placenta may have a direct adverse effect on the fetus.

The most important surgical risk to the fetus is preterm delivery. Hypoxia, infection, or genital tract manipulation (corpus luteum resection) may stimulate preterm labor. Before 23 weeks, preterm delivery uniformly results in neonatal death. When delivery occurs at a hospital with a level 3 perinatal center, survival at 26 weeks and 30 weeks is 70% and 90%, respectively. More than three quarters of neonates born after 28 weeks and 95% after 32 weeks at these centers will be free of major long-term handicaps. Birth at hospitals without expertise in high-risk obstetrics and neonatology is associated with a significantly higher risk of perinatal death and long-term handicaps.

An understanding of normal maternal-fetal physiology is critical in the diagnosis, surgical management, and postoperative care of pregnant women who require major surgery or who have been injured. Normal clinical and laboratory findings among pregnant women would suggest disease among women who were not pregnant. The differentiation between normal physiology and disease in these groups of women will save them from exposure to needless, and sometimes dangerous, diagnostic testing. Table 2 outlines the differences in laboratory values between pregnant and nonpregnant women.

Table 2. Laboratory Values in Pregnancy

BUN, blood urea nitrogen.

In the course of anesthesia, surgery, and postoperative recovery, the obstetrician, surgeon, and anesthesiologist monitor and manipulate critical physiologic variables such as blood pressure, heart rate, cardiac output, hematocrit, Pao₂, and Pa co₂. Table 3, adapted from a study by Clark and colleagues,⁸ depicts normal cardiovascular parameters in pregnant women. In their classic study, 10 normal pregnant women consented to serial placement of a Swan-Ganz catheter and measurement of their cardiovascular parameters in a resting, nonanesthetized state. These results are the values against which the values of the critically ill gravida should be compared.

Table 3. Central Hemodynamics in Pregnancy

NS, not significant.
(Data from Clark SL, Cotton DB, Lee W et al: Central hemodynamic assessment of normal term pregnancy. Am J Obstet Gynecol 161:1439, 1989.)

The specific organ system changes that lead to the variation in laboratory tests and cardiovascular parameters are described briefly in the following sections. These sections focus on physiologic changes relevant to the pregnant patient who is undergoing surgery or who has had a physical trauma.

Respiratory Physiology

Hormonal and mechanical changes (e.g., enlarging uterus) combine to produce hyperventilation. Progesterone acts early to increase respiratory rate by 15%. Minute ventilation and tidal volume are increased 50% and 40%, respectively. The relative hyperventilation is reflected by hypocapnia(Pa co₂ = 28 to 32 mmHg). The anxiety and pain can cause a pregnant patient to increase her respiratory rate to the point of significant hypocapnia, which results in faintness and perioral numbness. A 5-minute episode of hyperventilation that drops the maternal Pa co₂ by 6 mmHg results in a 4-mmHg drop in fetal Pa co₂ and, more importantly, a 3.5-mmHg drop in fetal Pao₂. Baseline hyperventilation also reduces respiratory compensation for metabolic acidosis.⁹

Inhalation anesthesia has a more rapid onset in pregnant patients. After the first trimester, progressive anatomic changes reduce the functional residual capacity and residual volume by 20%. When these changes are coupled with an increase in oxygen consumption (15%), hypoventilation or apnea results in a more rapid onset of hypoxia. During pregnancy, the closing capacity is elevated in the first 48 to 72 hours after abdominal or thoracic surgery; the resultant arteriovenous shunting further hastens the onset of hypoxia.

Cardiovascular Physiology

Hypervolemia, increased cardiac output, and positional hypotension are the major changes and risks of pregnancy-induced adaptations of the cardiovascular system. In normal pregnancies, oxygen consumption (o₂) is increased by a combination of increasing fetal metabolic requirements, maternal vasodilation, and arteriovenous shunting through the placenta. This metabolic demand is compensated by an increase in erythrocyte mass and cardiac index (CI). By the third trimester, there is a 50% increase in circulating blood volume. The plasma volume increases before the erythrocyte volume; it is normal for the hematocrit to decrease in the second trimester. A hematocrit of less than 30% is a diagnostic finding of anemia that is significant in pregnancy.

The full effects of vasodilation and arteriovenous shunting reach their maximum between 24 and 30 weeks, and a 30% reduction in both systolic and diastolic blood pressures is expected. Cardiac output rises rapidly until 30 to 32 weeks' gestation. This is accomplished by an increase in baseline heart rate to 90 beats/min and a lesser increase in stroke volume. In the third trimester, maternal position has a great impact on intravascular pressures and cardiac output. In the supine position, the uterus obstructs both the inferior vena cava and the aorta. Venous pressures in the lower leg can be as high as 20 to 25 mmHg, and the caliber of the aortic lumen is reduced by 35% to 40%. Cardiac output in the supine position subsequently may be reduced as much as 20% to 30% compared with that in the left recumbent position. Approximately 10% of pregnant women in the third trimester have supine hypotensive syndrome, which is manifested by profound maternal hypotension and fetal hypoxia.

The maternal cardiovascular system plays a key role in uteroplacental physiology. Familiarity with the maternal cardiovascular response to surgical hemorrhage, trauma, and sepsis is critical in understanding the complications resulting from these abdominal disorders and their treatment (e.g., surgery, vasoactive drugs). Circulatory physiology in response to challenge in nonpregnant animals and humans has been well described.¹⁰ In summary, tissue perfusion and aerobic tissue metabolism are a reflection of overall circulatory function. Oxygen delivery (Do₂), defined as arterial oxygen content multiplied by CI, reflects circulatory function. o₂, defined as arteriovenous oxygen content difference multiplied by CI, reflects the sum of all oxidative metabolic reactions plus arteriovenous shunting. The accumulation of metabolic acids, especially lactate, indicates a failure of the circulatory system (Do₂) to meet the oxygen needs (o₂) of the tissue; increased anaerobic metabolism is required. Patients with shock associated with hemorrhage or cardiac dysfunction have greatly reduced Do₂ with moderately reduced o₂. Increased oxygen extraction is the principal compensatory reaction to lowered o₂ when limited by Do₂. In contrast, increased o₂, which is provoked by increased metabolism from sepsis and trauma, is compensated by increased CI and Do₂. Therapeutically, management is directed at maintaining normal o₂ and Do₂ rather than focusing on a specific cardiac profile (e.g., blood pressure, wedge pressure). The way uterine arteries (i.e., uterine blood flow) participate in the maternal compensatory response to hypovolemia and shock reflect fetal tolerance. Knowledge of uterine blood flow and response to hemorrhage, hypoxia, or vasoactive drugs is derived from data in anesthetized animal models, usually ovine.^5,11–13a The data from animal experiments imply that the uterine vascular bed in pregnant organisms is a low-resistance system. The uterine vessels appear to be maximally dilated and therefore exhibit minimal autoregulation. More important, uterine blood flow reflects maternal arterial blood pressure in a linear fashion in pregnant organisms. When maternal blood volume is reduced by 30% to 35%, there is minimal change in maternal mean arterial pressure. As participants in the compensatory response, however, the uterine arteries constrict, and uterine blood flow is reduced by 10% to 20%.^11,12,13 α-Adrenergic compensatory vasoconstriction occurs in response to maternal hypoxia.^5,13 α-Adrenergic blocking agents, such as phenoxybenzamine, limit this response.^13,14 A decrease in uterine Pao₂ from 70 to 55 mmHg was associated in one study with a 16% reduction in uterine blood flow; a decrease in Pao₂ from 96 to 28 mmHg resulted in a 25% reduction.⁵

These observations contribute considerable interpretative information to clinicians faced with pregnant women with an acute abdomen. Clinical evidence of α-adrenergic stimulation raises a concern regarding reduced uterine blood flow and subsequent fetal compromise, despite normal maternal blood pressures. α-Adrenergic stimulation is manifested by marked peripheral vasoconstriction, cold and clammy extremities, a decrease in capillary refill in response to skin pressure, widened pulse pressure, decreased urinary output, and increased sweating.

Renal Physiology

Pregnancy-induced changes of the renal system include partial ureteral obstruction in the third trimester and increased glomerular clearance. A combination of smooth muscle relaxation by progesterone and compression by the enlarging uterus creates a physiologic hydronephrosis of pregnancy. Changes can occur before 12 weeks, and the obstruction is more pronounced on the right side. The anatomic changes revert to normal by 6 weeks postpartum. Changes in the cardiovascular system are reflected in the renal system. The 50% increase in blood volume during pregnancy results in an increased renal plasma flow and an increased glomerular filtration rate. As a result, there is an increase in the excretion of metabolic products (proteins and glucose) that may exceed the tubular reabsorption capability. In normal pregnancy, maximum blood urea nitrogen and creatinine levels are 10 mg/dL and 0.7 mg/dL, respectively. Glucosuria is found on two or more occasions in 5% to 10% of women as a result of a reduction in the renal threshold to between 140 and 160 mg/dL. Proteinuria levels of up to 300 mg/day are considered normal.

Coagulation Physiology

Thrombosis is more likely to occur in pregnancy because pregnant women have two of the components of Virchow's triad: venous stasis and hypercoagulability. The enlarged uterus compresses the vena cava and triples the venous pressure in the lower extremities (8 to 24 mmHg). Estrogen increases the hepatic production of coagulation factors, yielding a 30% to 50% increase in fibrinogen and factors VII, VIII, IX, and X. Fibrinolysis is decreased in the second and third trimesters. These changes increase the risk of thrombophlebitis to 1:70. Infection, general anesthesia, and surgery further exaggerate the risk when the third component of Virchow's triad, vascular injury, is added. Preoperative or postoperative bed rest further increases the risk of stasis. Prevention is the cornerstone of management. Once high-risk patients are identified, prophylactic heparin, intermittent compression devices, and early ambulation have significant benefit in reducing thrombosis and pulmonary embolism.

Gastrointestinal Physiology

Nausea and vomiting are common in pregnancy. In the first half of pregnancy, as many as 70% of pregnant women experience nausea and 40% report episodes of vomiting. The incidence is considerably reduced in the second half of pregnancy, but 15% to 20% of patients report persistence of symptoms throughout their pregnancy.¹⁵ The diagnosis of nausea and vomiting in pregnancy (hyperemesis gravidarum) is a diagnosis of exclusion. Any patient with nausea and vomiting deserves that a history be taken and a physical examination performed in a timely and thorough manner. Gastroenteritis is the most common nonobstetric diagnosis, but hepatitis, pancreatitis, and pyelonephritis also are frequent. Nausea and vomiting may be a component of appendicitis, but anorexia and pain usually are prominent. In a patient with previous pelvic or abdominal surgery, pelvic inflammatory disease, or endometriosis, intestinal obstruction is a consideration.

The physical examination is important in diagnosis and management. Abdominal distention, with or without ascitic fluid, is best determined by the location of areas of dullness when the patient is turned from side to side. Peristaltic sounds characteristic of an obstruction are high pitched, tinkling, and with rushes. Absence of peristalsis sounds is an important sign. Silence indicates ileus or the presence of intraperitoneal fluid, pus, or blood. Between these two extremes, wide variations in the type and quality of peristaltic sounds may be noted, but these differences usually are not important. When intestinal obstruction is suspected, abdominal x-ray films (upright, flat plate, and left lateral decubitus) are essential and should not be withheld for fear of fetal exposure to radiation.

Constipation is common in pregnancy because of the effects of progesterone on bowel mobility and the widespread use of hematinic agents; however, flatus always should be present. Anorexia, constipation, and the failure to pass flatus are serious signs that could indicate appendicitis, bowel perforation, or intestinal obstruction. Diarrhea seldom indicates an acute surgical problem in the abdomen except as a symptom of recurrent ulcerative colitis.

Physiologic changes in the gastrointestinal tract during pregnancy increase the risk of aspiration during surgery and anesthesia. Progesterone is a smooth muscle relaxant, and lower esophageal sphincter tone is decreased. One fourth of pregnant women have symptomatic lower esophageal reflux. Gastrin, the hormone that increases the volume of gastric secretion and lowers the pH of the stomach contents, increases considerably during pregnancy. In addition, intragastric pressure is increased by the large uterus, supine/lithotomy position, fundal pressure, or light anesthesia. Classic training has taught that approximately one fourth of gravidas undergoing abdominal surgery after an overnight fast have gastric contents of sufficient volume and of sufficiently low pH to place them at high risk of aspiration. Recent evidence demonstrates that pregnancy does not delay gastric emptying.^16,17,18 Despite the latter finding, an 8-hour fast, metoclopramide, and an H₂ antagonist should be prescribed as an antacid prophylaxis befor surgery.

Central Nervous System

Pregnancy changes the need for several anesthetic drugs. The requirement for halogenated anesthetics (halothane) is reduced by 40% because of the increased minute volume. Serum pseudocholinesterase is decreased by 20% in late pregnancy; however, clinically significant increases in the duration of succinylcholine neuromuscular block are reported rarely. Venous distention increases the caliber of epidural veins. Thus, the dose of epidural anesthetic agents is reduced by 50%. The uterine vasculature is more sensitive to α-adrenergic stimulation in pregnancy. Maternal stress response, maternal anxiety, or α-adrenergic vasopressors can result in fetal hypoxia and acidosis.

Acute Abdomen

Pregnant women are not immune to the varied conditions that can cause an acute abdomen in nonpregnant women. Some diseases are more likely to manifest during pregnancy; however, the diagnosis and management of an acute abdomen in a pregnant patient are modified in at least three ways. First, pregnancy induces physiologic and anatomic changes, and these changes require adjustments in diagnostic and therapeutic maneuvers. Pregnant women with an acute abdomen often come in contact with health care providers who are not trained in obstetrics and whose ignorance may lead to misdiagnosis, unwarranted caution, or ill-advised treatment. Conversely, some obstetricians may have limited or sporadic experiences with many medical and surgical diseases that occur occasionally in pregnant women. A delay in diagnosis and suboptimal or antiquated care can present a real danger to the patient. Second, complications of pregnancy extend the list of differential diagnoses of an acute abdomen (see later). Because complications of pregnancy may affect a fetus directly, a timely differentiation between nonobstetric and obstetric disease is critical. Third, a fetus is a passive participant in the disease processes. The disease process or management may directly affect a fetus through changes in uteroplacental physiologic makeup or by transport of bioactive agents across the placenta. Indirectly, a fetus may be affected by the initiation of preterm labor. After 20 weeks' gestation, the greatest risk to a fetus is preterm delivery.

The differential diagnosis of acute abdomen in pregnant women is as extensive as it is in all humans, but fortunately, ordinary ailments occur commonly. The causes of an acute abdomen in pregnancy fall into the following categories: extraperitoneal conditions, intra-abdominal/nonobstetric conditions, early obstetric complications, and late obstetric complications. Extraperitoneal conditions may involve referred abdominal pain, and these conditions include urinary tract infection, diabetic ketoacidosis, and sickle cell crisis. The major focus of this section will be intra-abdominal/nonobstetric conditions. The differential diagnosis includes viral gastroenteritis, appendicitis, cholecystitis/cholecystolithiasis, intrinsic bowel disease (regional enteritis/ulcerative colitis), hepatitis, pancreatitis, ovarian torsion, rupture of a corpus luteum cyst, infarction of a leiomyoma, or salpingitis. Other rare causes include intestinal obstruction, bowel perforation, diverticulitis, or peptic ulcer disease. In an increasingly mobile and perhaps more violent society, trauma is a common occurrence (6% to 7% of pregnancies). Potential abdominal injury usually is obvious in pregnant women with trauma. In domestic violence, the uterus and fetus may be the focus of the attack.

The incidence of most causes of an acute abdomen (e.g., appendicitis, gastroenteritis, intrinsic bowel disease, hepatitis) is unchanged by the presence of pregnancy. Some causes are less likely in pregnancy, such as diverticulitis, peptic ulcer disease, and salpingitis. Conversely, the following diseases are more likely in pregnant women than nonpregnant women: cholelithiasis/cholecystitis, ovarian torsion, infarction of a leiomyoma, rupture of a corpus luteum cyst, pancreatitis, urinary tract infections, and intestinal obstruction.

Before 20 weeks' gestation, the differential diagnosis of an acute abdomen is more difficult because more complications must be considered. These include early pregnancy complications: threatened abortion, septic abortion, ectopic pregnancy, salpingitis, ovarian torsion, and rupture of a corpus luteum cyst. When a normal uterus becomes an abdominal organ rather than a pelvic organ (at 16 weeks), acute abdominal symptoms may relate to an incarcerated uterus or bowel obstruction.

Later in pregnancy, the most common, confusing obstetric diagnoses are preterm labor, intra-amniotic infection, or abruptio placentae. Unfortunately, these complications may be the consequence, rather than the cause, of the acute abdomen. The differentiation between nonobstetric and obstetric causes is critical to management. The most effective aid in diagnosis is an experienced obstetrician who has conducted a detailed history and physical examination. Subsequently, the results of serial pelvic examinations, electronic fetal monitoring, and ultrasonographic examinations are important in the diagnosis and management. In selected cases, amniocentesis may provide useful information concerning amniotic fluid character (bloody or meconium stained), infection (Gram's stain and culture), and fetal lung maturity studies. When ultrasonography is not helpful, x-ray examinations should be performed in a calculated and focused manner. The fetal risks of diagnostic radiographs are minimal compared with the potential risk associated with delayed diagnosis or misdiagnosis of an acute abdomen.

Pain and associated abdominal tenderness are the most common and important symptom and sign. When pain is present, careful analysis of the location, type, intensity, and radiation is essential. Pain may be generalized or may be confined to the lower, middle, or upper abdomen. Generalized pain, when associated with rebound tenderness and guarding, is suggestive of severe irritation of the peritoneum, as found in pancreatitis, ruptured viscus, severe infection, or intra-abdominal hemorrhage. The presence of generalized pain and rebound tenderness subsequent to pain localized to one part of the abdomen suggests relatively sudden rupture of an organ, cyst, abscess, or tumor. With the rare exception of pancreatitis, early surgical exploration may be life saving for a mother and fetus.

Central lower abdominal pain suggests a uterine origin. Uterine contractions as in threatened abortion or preterm labor are the most common causes. Preterm labor may be the source or a consequence of the acute process. In either case, signs and symptoms of preterm labor must be evaluated. Preterm labor often is associated with one or more of the following signs or symptoms: waxing and waning central abdominal pain (like menstrual cramps), pelvic fullness, back pain, anterior thigh pain, groin pain, change in vaginal discharge, tenderness, urinary urgency, and vaginal bleeding or “spotting.” The diagnosis of preterm labor is based on progressive effacement and dilation of the cervix. Steady central lower abdominal pain is suggestive of an accident related to the uterus, such as uterine rupture, torsion or infarction of a leiomyoma, abruptio placentae, or chorioamnionitis, each of which may be accompanied by localized tenderness or rebound tenderness. In rare cases, urinary disorders (e.g., cystitis)can cause pain levels sufficient to be confused with more serious and acute intra-abdominal problems. Early in pregnancy, central abdominal pain may indicate an intestinal origin. If the pain is crampy, it suggests either enteritis or obstruction, which may result from adhesions or volvulus. Steady pain suggests inflammation such as appendicitis, mesenteric adenitis, Meckel's diverticulitis, regional enteritis, or, rarely, vascular accidents such as mesenteric thrombosis or embolism.

Lateralized lower abdominal pain suggests an adnexal problem such as infection, torsion, rupture, or hemorrhage. The pain caused by ovarian disease often radiates to the anterior thigh. Extraperitoneal causes of lateralized pain include the intense, colicky pain of a ureteral stone, which may radiate to the groin or labia. The pain of pyelonephritis may be abdominal but more often is prominent in the flank area. Intraperitoneal causes include appendicitis on the right side; the pain may be higher than usual because of the upward displacement of the appendix by the growing uterus (see later). As a result of displacement, the classic spread of pain from the umbilical area to the right lower quadrant of the abdomen may be absent in appendicitis. Other rare causes of right lower abdominal pain include regional enteritis, diverticulitis, and volvulus. In the left lower abdomen, pain may result from disorders of the descending and sigmoid colon such as ulcerative colitis, diverticulitis, or volvulus of the sigmoid.

Upper abdominal pain suggests disorders of the liver, gallbladder, stomach, duodenum, pancreas, or spleen. Gallbladder pain characteristically radiates to the back and to the right shoulder, whereas pancreatic pain is felt in the back. Splenic pain—most often from rupture and associated with other generalized and abdominal symptoms and signs—typically is referred to the left shoulder.

Appendicitis

In women, approximately 70% of cases of appendicitis involve women who are younger than 35 years of age; thus, the peak ages of childbearing and appendicitis coincide. Appendicitis occurs once in 750 to 1500 deliveries, but the incidence does not seem to be increased in pregnancy. Appendicitis accounts for two thirds of the operations on the gastrointestinal tract during pregnancy; biliary surgery and trauma account for an additional 24%. Unfortunately, the diagnosis of appendicitis is more difficult in pregnant than in nonpregnant patients.

Several factors contribute to the difficulty in diagnosis. Potential obstetric complications can mimic appendicitis, and changes in maternal anatomy and physiology obscure the diagnosis. In the first trimester, patients having spontaneous abortion, septic abortion, ectopic pregnancy, or ovarian accidents may present signs and symptoms similar to appendicitis. In late pregnancy, chorioamnionitis, abruptio placentae, or infarction of a leiomyoma can confuse the differential diagnosis. A second factor is the change in the position of the appendix. As the uterus enlarges, the appendix is pushed upward, laterally, and posteriorly. By the eighth month, 90% of appendices are 2 cm or more above the iliac crest and 3 cm lateral to McBurney's point. Although the initial visceral pain still is referred to the umbilicus, the posterior displacement of the appendix in late pregnancy eliminates the localized peritonitis described classically in nonpregnant patients. Weingold¹⁹ cataloged the changes in the physical findings related to gestational age and appendicitis. Rebound tenderness was present in 100% versus 43% of patients in the first versus third trimester, respectively. Rectal tenderness was found in 67% of cases of appendicitis in the first trimester and in no cases of appendicitis in the third trimester. The incidence of right lower quadrant tenderness was 100% in appendicitis cases in the first trimester and 14% in those in the third trimester. Right upper quadrant tenderness occurred in no appendicitis cases in the first trimester but in 60% of those in the third trimester. The latter changes probably account for the higher incidence of perforated appendices in late pregnancy: 30% in the first trimester and 70% in the third trimester.

Once the diagnosis is suspected, surgery should not be delayed: The incidence of perforation increases from 20% in the first 24 hours of symptoms to 70% after 48 hours of symptoms. When in doubt about the diagnosis, the surgeon should make a midline incision, which allows greater exposure to the left adnexa. In the first trimester, diagnostic laparoscopy is invaluable. When the diagnosis is clear in the first trimester, a transverse incision is made 1 to 3 cm below the umbilicus and is centered on the midclavicular line. Later in pregnancy, the transverse incision is centered on the point of maximum tenderness, and thus is moved more cephalad and laterally as the uterus enlarges. The length of the incision is about the width of the surgeon's hand. The aponeurosis and muscles are incised or split in the direction of their muscle fibers. The anterior rectus sheath, but not the enclosed muscle, is incised to allow better retraction. The inferior epigastric artery, which lies on the posterior rectus sheath, must be carefully avoided or ligated. In most women in the first trimester, the appendix can be seen readily at the incision site. The anterior cecal taenia and the ileocecal junction are helpful anatomic landmarks. Extensive exploration should be avoided because necrotic, friable tissue may be involved. The mesoappendix is progressively clamped, cut, and ligated with hemostats, Metzenbaum scissors, and 00 polyglycolic sutures. In most cases, this step proceeds distally to proximally. In some cases of retrocecal appendices, the direction is reversed. The free appendix is crush-clamped at its base and clamped distally, and a single 00 plain catgut ligature is placed in the groove. A second ligature is used if needed for control. The appendix is excised close to the clamp, and the knife and clamp are removed from the field. The use of either phenol and alcohol or electrocautery on the stump probably does not affect outcome. However, if it is not neutralized with alcohol, phenol can burn normal bowel on contact. The stump is inverted with 000 silk purse string or Z-type sutures. The advantages of this technique are security of the intramural branch of the appendiceal artery (ligation) and reperitonealization (inversion).

The appendiceal bed is irrigated with saline, and purulent debris is removed. The use of antibiotics in the irrigant may be beneficial in preventing abscess formation. The abdominal wall is closed in the usual fashion. In cases of gross contamination, the skin is left open for delayed closure. Although not always indicated in uncomplicated appendicitis in nonpregnant patients, prophylactic intravenous antibiotics are recommended in pregnant patients. A second- or third-generation cephalosporin with anaerobic and gram-negative coverage is an appropriate choice. Antibiotics should be continued until a patient's temperature has been less than 37.4°C for 48 hours.

Perforation occurs in 30% to 70% of cases of appendicitis complicating pregnancy. Perforation with abscess formation demands different intraoperative management during pregnancy. In nonpregnant patients with a fixed periappendiceal mass, expectant management with observation and intravenous antibiotics may be advocated. Appendectomy is performed at a later date. However, pregnant women are less able to contain the infection, and the risk to the fetus from endotoxins is high. Surgical management always is indicated in pregnancy.

A retroperitoneal approach is used to avoid additional peritoneal contamination. The location of the abscess usually corresponds to that of the tip of the appendix. A transverse muscle-splitting incision is made just medial to the crest of the ilium. The lateral edge of the peritoneum is exposed, and the abscess is entered from its lateral and retroperitoneal aspect. Aerobic and anaerobic cultures are taken, and locations are broken down by gentle finger dissection. The adhesions walling off the abscess should not be disrupted to avoid further peritoneal spread. If the appendix is readily accessible, an appendectomy is performed. A closed-suction drainage system then is brought out through a separate flank incision (inside to outside). The wound and abscess cavity are irrigated with copious amounts of diluted antibiotic solution. The fascia is closed with permanent sutures (nylon) using a Smead-Jones technique. The skin is packed open for delayed closure. The closed drainage system should be allowed to drain undisturbed for 24 to 48 hours, after which, if it is draining less than 100 mL/24 hours, it may be rotated and removed gradually during a period of 2 to 4 days.

Occasionally, pregnant women present with diffuse peritoneal signs. When diffuse peritonitis with fever is present, rapid surgical exploration is indicated; often, a ruptured appendix is the source. A midline abdominal incision is used, allowing greater exposure for diagnosis and management of intra-abdominal abscesses. After 20 weeks' gestation, an upper midline incision is required for exposure. The appendix is examined and removed in most cases. Localized collections of pus are drained by soft rubber drains. The use of prophylactic drains is not recommended because they are rapidly walled off, thus defeating their purpose. The abdominal cavity is irrigated with copious amounts of diluted antibiotic solution. The fascia is closed by permanent sutures using the Smead-Jones technique. In anticipation of wound infection, the skin and subcutaneous fat are packed with saline-moistened gauze. The dressing is changed three to four times a day, and if no infection is present 3 or 4 days after surgery, the skin is closed with the patient under local anesthesia.

Systemic intravenous antibiotics should be continued for a minimum of 5 days; after a 5-day course, antibiotics can be discontinued when the patient has been afebrile for 24 hours. The head of the bed should be elevated 15 to 30 degrees in the first few days after operation. Although this does not prevent subphrenic abscess, it promotes drainage toward the pelvis.

Complications are relatively common: 5% to 10% in early, nonperforated appendicitis; 30% to 50% when perforation has occurred.²⁰ Maternal mortality is rare but possible (0.1% to 0.5%). The chief complications are wound infections, abscesses, and peritonitis. Continued fever suggests further abscess formation. Rectal examination may detect a developing pelvic abscess. Vaginal examination should be used only for detection of an obstetric complication (e.g. preterm labor). Preterm delivery always is a danger, occurring in 4.5% without perforation and in 27% with perforation.

Cholecystitis

Cholecystectomy is the second most common nonobstetric operation performed during pregnancy (1 to 5/10,000 pregnancies). The reasons are an increase in cholesterol synthesis, stasis of bile in the gallbladder, increased concentration of cholesterol in bile, and decrease in the circulating bile salt pool.

Cholecystitis (cholelithiasis with infection) is the major indication for cholecystectomy during pregnancy. The mere presence of stones without infection, even if symptoms are present, is not an indication for operation; it should be deferred until after pregnancy. The clinical features of cholecystitis are similar to those in nonpregnant patients. Symptoms consist of right upper quadrant abdominal or epigastric pain that radiates to the back or right shoulder, flatulence, heartburn, nausea and vomiting, and fatty food intolerance. Physical findings include fever, right upper quadrant abdominal tenderness, and leukocytosis (more than 16,000/mm³). Jaundice may be present if the common duct is obstructed, but other causes of jaundice need to be ruled out with appropriate tests. Hyperamylasemia is present in many patients but usually subsides with hydration. The definitive diagnosis of gallstones is made by ultrasonography. Preoperative cholecystography and radionuclide scanning should not be performed in pregnancy because of possible radiation danger to the fetus.

The initial treatment of acute cholecystitis is conservative²⁰: nasogastric suction, intravenous fluids, and rest. Morphine is contraindicated because it can exacerbate colic. Broad-spectrum antibiotics (e.g., second-generation cephalosporins) probably are useful. Acute symptoms subside in approximately 80% of patients. If the symptoms do not subside or if signs of sepsis, peritonitis, or increasing jaundice supervene, cholecystectomy is indicated. Surgical exploration of the common duct should not be withheld because of pregnancy, but operative cholangiography or ultrasonography can provide a valuable means of avoiding such a significant additional procedure. The results of cholecystectomy usually are good for the mother, but pregnancy loss is higher after cholecystectomy than as a result of other causes of acute abdomen.²¹ As is true of all other abdominal operations during pregnancy, careful monitoring of the fetus after operation is essential.

Laparoscopic cholecystectomy is being advocated for treatment of cholelithiasis. More surgeons are performing laparoscopic cholecystectomy during pregnancy. Few data support any differences in safety between open and laparoscopic cholecystectomy. Overall, operative laparoscopy is associated with 10.4% complications, and half of these are major.²² One study compares open (n = 26) versus laparoscopic cholecystectomy (n = 20) among pregnant women. Among the women who had laparoscopic cholecystectomy, one patient and her fetus died from hemorrhage. In the traditional, open cholecystectomy group, one fetus died 5 weeks after surgery.²³ Animal studies demonstrate fetal acidosis with CO₂ pneumoperitoneum. The fetal acidosis may explain four fetal deaths in seven cases of operative laparoscopy during pregnancy.²⁴

The second major complication is inadvertent trocar injury to the enlarged uterus. I have encountered a patient with a life-threatening injury to the uterus when a surgeon attempted a laparoscopic cholecystectomy at 30 weeks. Whereas the safety will increase with more experience with operative laparoscopy during pregnancy, the procedure has major morbidity and should be limited to failed conservative therapy before 16 weeks' gestation.

Intestinal Obstruction

Intestinal obstruction in pregnancy is more common than it was 30 years ago. It is estimated to occur in 1 of 3000 deliveries. The increase probably results from more abdominal operations (and subsequent adhesions) and a higher prevalence of pelvic inflammatory disease among young women. Obstruction seems to occur most frequently when the uterus becomes a suprapelvic organ at 12 to 14 weeks' gestation, when the head of the fetus descends into the pelvis late in pregnancy or in labor; or in the early puerperium, or when rapid emptying of the uterus changes the relationship of bowel to fixed adhesions. As in nonpregnant patients, intestinal obstruction in pregnant patients can be partial or complete, or it can be simple or strangulating. Strangulation occurs when bowel vasculature is compromised by thrombosis or obstruction. The typical symptoms of intestinal obstruction—crampy abdominal pain, distention, failure to pass gas and stool, and nausea and vomiting—sometimes are difficult to distinguish from those of normal pregnancy. Nausea and vomiting occur frequently in early pregnancy, and crampy uterine contractions are common in later pregnancy. Physical examination typically shows some abdominal distention and hyperactive peristalsis with rushes. Fever, abdominal tenderness, leukocytosis, or electrolyte abnormalities increase the possibility of strangulation. If the initial picture is not clear, it may become more obvious with regular observation, during which the fetus must be monitored regularly. Serial flat and erect x-ray films of the abdomen are necessary and justifiable.

Conservative management includes nasogastric suction (short or long tube), intravenous fluid replacement, and parenteral nutrition. Because a delay in definitive therapy (surgery) is associated with increased fetal risk, early surgical treatment is recommended. If the diagnosis is reasonably well established, if strangulation is suspected, or, if symptoms do not improve with conservative measures, prompt surgical exploration is preformed. At operation, a long midline incision should be used and adhesions divided or nonviable bowel resected if necessary. The operative mortality rate for intestinal obstruction in pregnancy is 5% to 15%, mostly as a result of delay in definitive therapy.

Other Intestinal Disorders

Although gallbladder disease is the most common upper abdominal disorder in pregnancy, the differential diagnosis includes pancreatitis and perforated peptic ulcer. Pancreatitis is one of the hazards in women who have greatly elevated triglycerides in pregnancy (above 500 mg/dL). The hyperamylasemia of cholecystitis may be confused with that of pancreatitis, but the former improves with hydration. Perforation of an ulcer results in peritoneal irritation similar to that encountered with a ruptured ectopic pregnancy, but without hemorrhagic shock. Erect abdominal films usually show air under the diaphragm.

Other intestinal disorders that can occur in pregnancy include Meckel's diverticulitis, inflammatory bowel disease (which may improve during pregnancy because of increased steroid production), and diverticulitis of the colon. The significance of all of these conditions and other more uncommon conditions lies in an awareness of the possibility of their occurring. Once the diagnosis is made, treatment resembles that for nonpregnant women. The caveats presented here about the danger and difficulty of any operation in pregnant patients, especially those in the third trimester, and the risk of premature labor are intended to encourage a persistent conservative approach.

Disorders of the ovaries, tubes, and uterus account for a large proportion of the acute abdominal conditions encountered in pregnancy. These may be either related or unrelated to the pregnancy itself. Symptoms and signs typically involve the pelvis but may be generalized to the whole abdomen.

Ovarian Disease

The ovaries are the sites of a wide variety of cysts and tumors, benign and malignant, found in pregnant women. Routine ultrasonography at 12 to 18 weeks demonstrates ovarian masses in 0.6% (5 to 6 per 1000) of pregnancies.²⁵ The most common size is 7 to 11 cm, and 27% are larger than 10.9 cm. Table 4 depicts the histologic type of ovarian tumors in pregnancy.^26,27,28 The presence of an enlarged ovary in a pregnant woman deserves the same careful investigation as in a nonpregnant woman. Although removal may be indicated on the basis of size or possible malignancy, it is usually not an urgent matter. Emergencies include ovarian rupture, hemorrhage, tumor, or torsion. Pain is the most common symptom (48%). A ruptured ovarian cyst causes generalized pain and peritoneal irritation. Hemorrhage within a cyst and torsion of a cyst or tumor cause pain that is localized laterally in the lower abdomen. Fever and leukocytosis accompanies torsion if the tumor has become necrotic. Torsion seems to occur more often in pregnancy: 28.6% versus 7.3% in nonpregnant patients.²⁷

Table 4. Histologic Type of Ovarian Tumors in Pregnancy

Diagnostic studies including ultrasonography can be useful in identifying a cystic or solid mass as distinct from the uterus, whereas blood obtained by culdocentesis or paracentesis is diagnostic of intraperitoneal hemorrhage than the patient's hematocrit. X-ray films of the abdomen or pelvis are rarely helpful and involve unnecessary radiation exposure to a pregnant woman and her fetus. Laparoscopy is a valuable diagnostic aid early in pregnancy when (1) symptoms and signs of an accident are equivocal; (2) the ovary is small (less than 6 cm); or (3) it is not clear whether the tube, ovary, or another organ is the primary site of the problem.

A bleeding corpus luteum cyst is a major differential diagnosis in the evaluation of a patient in early pregnancy that has abdominal pain. The major differential diagnosis is ectopic pregnancy; in both disorders, the patient presents with lower abdominal pain and tenderness. In addition to pregnancy testing and ultrasonography, culdocentesis can be helpful. Culdocentesis with the finding of a hematocrit less than 15% usually is associated with a minimal amount of ovarian bleeding. Most of the fluids from ruptured ectopic pregnancies have hematocrits above 15%. When the hematocrit of the peritoneal fluid is above 15%, diagnostic laparoscopy usually is indicated. The corpus luteum is essential for hormonal (progesterone) support of the pregnancy until the placenta can support itself. The corpus luteum is indispensable until about 9 weeks' gestation. Lute- ectomy before 9 weeks may result in spontaneous abortion; however, luteectomy occasionally is necessary, and in these patients, 200 mg of progesterone intramuscularly each day is essential for pregnancy support until 9 weeks' gestation.

After it has been concluded that a significant ovarian accident has occurred, an expedient exploratory laparotomy should be performed. The cardinal principles of ovarian surgery apply. The obstetrician should make a vertical incision that can be extended upward. If any fluid is found on entering the peritoneal cavity, it should be sent for cytologic examination before any manipulations are to be performed. If no fluid is present, the pelvis should be irrigated with 100 to 200 mL of saline, and the irrigant should be sent for cytologic examination. If an ovarian cyst or tumor of uncertain nature is found, it should be excised and examined grossly and, if possible, microscopically for evidence of malignancy. An exception to removing the whole ovary may be the presence of a benign cystic teratoma or simple cyst. In this case, the cyst should be excised from the substance of the ovary, and the shell of ovarian tissue should be reconstructed. If the ovary or tube is twisted, the pedicle should be clamped first; it should not be untwisted, thus preventing the toxic products of necrosis or blood clots to enter the general circulation. In cases of a malignant or potentially malignant ovarian cyst or tumor, the abdomen, including the undersurface of the diaphragm and the para-aortic nodes, should be explored carefully. Additional histologic samples of the omentum, peritoneum of the diaphragm, gutters and pelvis, and any enlarged para-aortic nodes should be obtained. If a malignant tumor is found, the decision about how to proceed is difficult. Consultation should be obtained with a gynecologic oncologist, if possible. One or both ovaries can be removed after the first 10 weeks of pregnancy without affecting the hormonal support of the pregnancy.

Ectopic Pregnancy

The fallopian tube is the most common site of an ectopic pregnancy, and an intact or ruptured tubal pregnancy is responsible for most of the acute abdominal conditions arising in the pelvis during early pregnancy. The diagnosis often is established by the classic picture of ectopic pregnancy, vaginal bleeding, and lateralized lower abdominal pain (becoming generalized with rupture and occasionally associated with shoulder pain and fainting). It can be confirmed by the finding of a sausage-shaped tender tubal mass and nonclotting blood on culdocentesis. Vaginal or abdominal ultrasonography can rule out an ectopic pregnancy by demonstrating an intrauterine pregnancy, or it can diagnose an ectopic pregnancy by demonstrating an extrauterine gestational sac.

Laparoscopy is important in differentiating the causes of an acute abdomen in early pregnancy and treatment of many causes, including ectopic pregnancy. If an unruptured tubal ectopic pregnancy is diagnosed, laparoscopic removal is the surgery of choice. Laparotomy and salpingectomy should be reserved for a ruptured or large ectopic pregnancy. At laparotomy, the ovary should be left in place; other surgical maneuvers, such as appendectomy, should be avoided. Blood should be removed from the peritoneal cavity.

Leiomyoma

Leiomyoma of the pregnant uterus is a common occurrence (5% to 20%). Most leiomyomas are asymptomatic and remain undiagnosed unless they are larger than 6 cm. Aseptic necrosis, torsion of the pedicle, or parenchymal hemorrhage occur in 10% to 20% of pregnant patients with recognized leiomyomas.²⁰ These events manifest as acute abdominal pain or preterm labor. Additional complications include abnormal fetal presentation, dystocia, and obstetric hemorrhage (antepartum and postpartum). The diagnosis of abdominal pain related to a leiomyoma is difficult and requires elimination of other diagnoses. Accidents related to leiomyomas are treated conservatively with analgesics. If the diagnosis is made by laparoscopy or laparotomy, myomectomy should not be performed except in extreme circumstances and only if leiomyoma is attached to a small, accessible pedicle.

Cervical Incompetence

Cerclage for cervical incompetence is a common surgical procedure during pregnancy. Cervical incompetence is caused by trauma (40%), congenital abnormalities of the cervix (40%), and unknown causes (20%). Cervical incompetence accounts for 15% to 25% of midtrimester loss. Approximately 0.5% to 1% of all pregnancies are complicated by an incompetent cervix.²⁹ The McDonald or Shirodkar cerclage are the most common corrective procedures. There is an overall risk of 2% to 4% for a pregnancy loss with either procedure. The complications are listed in Table 5. The latter risks are increased with emergency cerclage, cervical dilation more than 2 cm, suture after 20 weeks, history of spontaneous abortion at less than 24 weeks, and vaginal infections(e.g., group B Streptococcus, bacterial vaginosis, Trichomonas infection, vaginitis, gonorrhea, Chlamydia infection). The clinical presentation at the time of the cerclage impacts the success of the procedure.³⁰ The average gestational age at delivery for elective cerclage is 34 to 36 weeks; for urgent cerclage (asymptomatic, shortened cervix, or funneled cervix), 31 to 34 weeks; and for emergency cerclage (symptomatic), 29 to 31 weeks. The incidence of premature rupture of membranes is 18%, 40%, and 58% in elective, urgent, and emergent cerclage, respectively.

Table 5. Complications of Cerclage

Cervical Neoplasia

Papanicolaou (Pap) smears during pregnancy have become a standard of care in the United States. As a consequence, pregnant women constitute the best-screened population for cervical dysplasia. In the last 20 years, the evidence of cervical intraepithelial neoplasia (CIN) has risen from 0.2% to 0.5% in the 1960s to between 3% and 5% in the early 1990s and 5% to 10% in 2000. Similarly, the incidence of carcinoma in situ (CIS) has risen from 1 : 1600 in the middle 1960s to 1:500 in the early 1990s. The incidence of invasive cervical carcinoma in pregnancy has stayed relatively constant at about 1 : 2000. The diagnosis and treatment of CIN in pregnancy involves surgical and pregnancy risks.

The decision-making process required for management of an abnormal Pap smear in pregnancy requires expertise and experience in colposcopy. In general, an obstetrician who performs more than five colposcopies a week has the interest, skill, and experience to manage CIN in pregnancy. During pregnancy, atypical squamous cells of undetermined origin and low-grade squamous intraepithelial lesions should be managed conservatively with treatment of vaginal infections and repeat Pap smear in 8 weeks. If the abnormality is persistent, the patient needs a colposcopic examination and directed biopsy. Hacker and coworkers³¹ reviewed nine studies of colposcopy in pregnancy, which involved a total of 1064 patients. The diagnostic accuracy was 99.5%, and no cases of invasive carcinoma were missed. The complication rate was 0.6% and consisted mostly of excessive bleeding at the biopsy site. Bleeding can be controlled with silver nitrate or ferrous subsulfate (Monsel's solution). Notice that endocervical curettage is not recommended during pregnancy because of potential damage to the fetus. Random biopsies or four-quadrant biopsies of the cervix should not be performed. Random biopsies had a false-negative rate of 8% to 40%—an accuracy far worse than colposcopically directed biopsy.

High-grade squamous intraepithelial lesions (HGSIL) require immediate colposcopic examination and directed biopsy. If results of the colposcopy are satisfactory and biopsy demonstrates HGSIL, a follow-up Pap smear and colposcopy should be planned every 8 weeks for the remainder of the pregnancy. A vaginal delivery and definitive postpartum therapy is expected. Histologically proven progression in pregnancy to a higher grade of dysplasia postpartum occurs in 5% to 10%.³² Among 153 patients with moderate or severe dysplasia diagnosed and not treated during pregnancy, 69% had regression of their dysplasia postpartum; 7% progressed from moderate dysplasia to severe dysplasia. Mode of delivery did not affect the progression rate.³³ No patients progressed to invasive disease. In 25 patients with CIS diagnosed during pregnancy but not treated, 80% had persistent CIS, 8% had progression to invasive disease postpartum, and 12% had resolution of the CIS without treatment at postpartum evaluation.³⁴ Colposcopy without biopsy underestimated the severity of the lesion in 35% of the cases; biopsy of the worst lesions is recommended.³⁴ A note of caution was given in the study by Woodrow and colleagues.³² Fifteen percent to 30% of patients with cervical dysplasia during pregnancy are lost to follow-up.^33,34 These data underline the importance of methods to maximize postpartum patient follow-up.

If microinvasion, adenocarcinoma, histologic discrepancy of Pap smear findings, or an unsatisfactory colposcopy result are found, the patient should have colposcopically directed excision of the affected area. In the first 32 weeks, this excision is performed immediately. In the absence of invasive carcinoma on subsequent histologic study, a vaginal delivery should be anticipated and definitive therapy planned for the postpartum period. If the diagnosis is made in the last 8 weeks of pregnancy, surgical excision may be delayed until 6 weeks after a vaginal delivery, unless invasive disease is suspected. If invasive carcinoma is diagnosed at any point in the workup, further diagnosis (e.g., magnetic resonance imaging, computed tomography [CT]) should be performed and therapy planned. Consultation with the gynecologic oncologist, maternal fetal medicine specialists, and family are recommended for planning the treatment of invasive cervical cancer during pregnancy.

The principle goal of cone biopsy during pregnancy is to rule out invasive disease, not to treat the lesion. Persistent disease and increased complications dictate this recommendation. Surgical excision of the dysplastic regions is a less effective treatment in pregnant versus nonpregnant women. Hacker and coworkers³¹ reviewed 376 cold-knife cone biopsy results of pregnant women and found that 43% of patients had residual neoplasms (CIS) on histologic examination in the postpartum period. Cytology and colposcopy should be repeated every 10 to 12 weeks postoperatively. Definitive therapy should be given 6 weeks into the postpartum period.

The risks of cold-knife conization during pregnancy are considerable. Cold-knife cone biopsy for CIS during pregnancy results in a 60% increase in preterm and low-birthweight infants after control for smoking, race, parity, marital status, and history of pregnancy terminations.³⁵ Newton²⁰ reviews the short-term complications of cold-knife cone biopsy in pregnancy. In five studies involving 448 pregnant women who underwent cold-knife cone biopsy during pregnancy, the following complications were found:

  Spontaneous abortion, 18% (10/56)
  Perinatal mortality rate, 4.1% (16/387 [41/1000 births])
  Immediate bleeding, 7.4% (33/448)
  Bleeding after 5 days, 3.3% (15/448)
  Transfusion, 7.4% (33/448)
  Premature delivery, 10.6% (41/387)
  Cervical laceration, 7.7% (27/351)

In summary, cold-knife cone biopsy during pregnancy is associated with significant maternal morbidity, a perinatal mortality rate four times the overall rate, and small but definite increases in second-trimester spontaneous abortions and preterm deliveries.

In the last 10 years, two new techniques have been developed to excise abnormal tissue from the cervix: loop electrosurgical excision procedure (LEEP) and laser conization. Neither technique has been studied adequately in pregnant women for its effect on pregnancy outcome; each has been studied mainly in relation to its effect on subsequent pregnancy. One descriptive study reports the outcome of 20 LEEP procedures during pregnancy.³⁶ Fifty-seven percent had cervical dysplasia at the margin; two patients required blood transfusions; 15% had preterm deliveries; and, one fetal death occurred 4 weeks after the procedure. No studies describe the outcomes of laser conization during pregnancy. Until more data are provided, the clinician must assume that neither LEEP or laser conization offer safety advantages over cold-knife conization during pregnancy.

Fortunately, frank invasive carcinoma of the cervix is rare: 1 of 2000 pregnancies. Approximately 3% of all cases of cervical carcinoma are associated with pregnancy. There is a significantly increased risk of finding a higher stage in late pregnancy, although there is an equal incidence of carcinoma in all trimesters. Approximately one half of pregnancy-associated carcinomas of the cervix are diagnosed in the postpartum period.³¹

Pregnancy complicates the diagnosis of cervical carcinoma. As with the nonpregnant patient, the most common symptom is painless bleeding. However, bleeding also is a prominent feature in many pregnancy complications, such as threatened abortion or placenta previa. Thus, diagnosis can be delayed in as many as 62% of cases of cervical carcinoma diagnosed during pregnancy.³¹ Invasive carcinoma may be present even with a recent negative finding on Pap smear. As many as 20% of screened patients who subsequently developed invasive carcinoma of the cervix had at least two negative Pap smear results within 3 years of the diagnosis of cancer.³¹

The management of carcinoma of the cervix in pregnancy is complicated by the presence of the fetus. In general, before 20 weeks, treatment should be begun promptly without regard to the fetus because delay until viability might permit the cancer to grow beyond easy control. At 20 to 34 weeks, some delay in therapy is appropriate to permit the fetus to mature, but each case requires individual consideration. After 34 weeks, the fetus should be delivered by cesarean section and definitive treatment given as soon as possible.

The principles of treatment are similar to those for the nonpregnant woman.^31,37,38,39 Although extrafascial hysterectomy is appropriate treatment for microinvasive (stage IA) lesions, total excision or conization may be acceptable during pregnancy in some instances if the lesion is completely excised. Radical hysterectomy is complicated by an enlarged uterus and increased blood supply to the pelvis. It is best to empty the uterus through a classic incision before the hysterectomy. Special care should be taken with hemostasis. If radiation therapy is to be used, abortion usually occurs spontaneously during external radiation therapy in the first or early second trimester. Hysterotomy before radiation therapy may avoid problems of managing complications such as abortion or intrauterine fetal death.

When inadvertent delivery occurs through a cervix with an invasive carcinoma there is no conclusive evidence that this worsens the prognosis for the disease³¹; however, there is a considerable risk of hemorrhage at the tumor site if the lesion is visible. As in the nonpregnant woman, the clinical stage of invasive carcinoma of the cervix is the most important determinant of prognosis. Pregnancy does not seem to affect prognosis in stages I or II. For more advanced disease, the prognosis appears to be unfavorably affected by pregnancy. The complexity of dosimetry in a pregnant uterus and the increased number of interruptions of the dose schedule for genital infections may be related to the poorer prognosis.^31,37,38 The prognosis for 5-year survival is 74.5% in stage IB, 47.8% in stage II, and 16.2% in stages III and IV.

Breast Mass

Enlargement of the breasts during pregnancy and the production of milk after delivery alter the clinical aspects of breast masses and their management. Fibrocystic changes become less obvious, although cysts occasionally are felt. Fibroadenomas may enlarge, sometimes to a considerable size (lactating fibroadenoma). A rare condition complicating pregnancy is gigantomastia: the size and weight of the breasts, as well as the ulceration of the dependent skin, can render the patient completely inactive and make clinical management difficult. Cancer can occur in women who are pregnant or lactating, especially in those who are 35 years of age or older. Its growth does not seem to be affected by pregnancy; results of treatment are similar, stage for stage, in pregnant and nonpregnant women.

The steps to be taken in the diagnosis of breast masses in pregnancy do not differ greatly from those in the nonpregnant woman. Physical examination is more difficult because of lactational growth (200 to 600 g per breast during pregnancy), and dominant masses may not be as easy to feel. Fine-needle aspiration with a no. 22 needle can be performed just as is done with other solid organs, without danger of excess bleeding or the formation or milk fistula. Mammography may be less helpful; the density of pregnant breast tissue makes interpretation more difficult. Ultrasonography may be more desirable than mammography as a screening test because if offers no danger to the fetus; however, its main value is to determine whether a given mass is cystic or solid. A cystic mass can be aspirated; a biopsy is necessary if the mass is noted to be solid on ultrasound examination or remains after needle aspiration. If a biopsy is necessary, a circumareolar incision may result in the division of the lactiferous ducts or injury to the nerves supplying the nipple. There is a threefold increase in lactational insufficiency after circumareolar incisions.⁴⁰ A lateral incision should be made parallel to the areola and the tissues spread in a radial direction. Local anesthesia should be used. Bleeding may be heavier than anticipated and should be controlled with individual ligatures to avoid mass compression of ducts. During lactation, mild stasis may predispose the patient to infection, and milk fistula can form at the site of the incision.

The definitive treatment of breast cancer in pregnant versus nonpregnant women generally does not differ; however, any operation chosen for a pregnant woman should lean on the side of conservative versus radical if there is no significant difference in long-term maternal survival. The longer and more extensive the operation is, the greater the risk to the fetus. Reconstruction of the breast may best be left until after the pregnancy is completed. The timing of postoperative therapeutic radiation or chemotherapy is governed by residual tumor, gestational age of the fetus, and actual fetal exposure to these therapies. The family must play an important part in any decision for operative and postoperative treatment.

Breast Infections

Breast infection (abscess) occurs at a baseline rate of 3% to 5% of puerperal women but may be more frequent in “epidemics.” It usually occurs within weeks after delivery but may occur at any time after that, with some increase at the time of weaning, especially if this is done quickly. The onset of infection occurs in two stages: first by an initial duct obstruction that presents as a tender nodule (blocked duct), followed by local or general mastitis. A blocked milk duct (the former) often subsides with adequate emptying, either by the infant or by a breast pump. The key signs of mastitis are a wedge-shaped area of tenderness, erythema, and induration. The patient seems toxic with a temperature of 38° to 40° (101° to 104°F). If emptying the breast is unsuccessful or if the patient initially presents with mastitis, antibiotic therapy is indicated to combat the most likely infectious agent, Staphylococcus aureus. At the same time, it is best for the infant to continue to breast-feed, first on the unaffected side to promote milk ejection and to permit easier breast-feeding from the affected side. A breast pump also may be used. Mastitis progresses to breast abscess in 3% to 5% of cases.

Aspiration of the abscess under ultrasonographic guidance and intravenous antibiotics (appropriate for infections with Staphylococcus aureus) are the standard management of a breast abscess. Eighty-six percent of women respond without surgery.⁴¹ Surgical drainage is indicated when fluctuation is apparent. The incision should be over the area of fluctuation. The cavity can be packed or drained and antibiotics continued. Abscesses recur in up to 25% of cases. Further drainage may be necessary, and in this case it may be advisable for the patient to discontinue breast-feeding on the affected side; milk production from the unaffected side increases to compensate for the loss of breast-milk volume from the affected side.

Trauma

Trauma caused by accidents and violence is a common and important complication of pregnancy, involving 5% to 20% of pregnancies.^20,42 Population studies^43,44 of injured patients reveal 34% related to transportation, 26% to falls, 16% to poisonings, and 16% to domestic violence. In 372 injured patients, 84% had blunt injuries and 16% had penetrating injuries.⁴³ The incidence of maternal death is 3% to 8%. Studies demonstrate that trauma is more likely to cause maternal death than any of the medical complications of pregnancy. The medical records of the Cook County Medical Examiner were reviewed for the years 1986 to 1989.⁴⁵ Direct and indirect obstetric factors caused maternal deaths in 31.5% of 95 cases. Trauma caused maternal deaths in 46.5% of the 95 cases; of these traumatic death cases, 34% were caused by accidents; 57%, homicide; and 9%, suicide. Table 6 describes the mechanism of injury in traumatic maternal death in this series.⁴⁵ The records of the New York City Medical Examiner from 1987 to 1991 identified women aged 15 to 44 years who were pregnant at the time of their traumatic death.⁴⁶ Homicide (63%), suicide (13%), motor vehicle accidents (12%), and drug overdoses (7%) accounted for the deaths; 48% of the injury deaths were associated with recent substance use.

Table 6. Mechanism of Injury in Traumatic Maternal Death

(Data from Fildes J, Reed L, Jones N et al: Trauma: The leading cause of maternal death. J Trauma 32:643, 1992.)

Fetal compromise roughly correlates with maternal condition at admission. When the patient is admitted with a trauma diagnosis, the incidence of preterm birth and abruption are 18% and 2%, respectively. An injury severity score above 25 is associated with 50% incidence of fetal death. Abruption is associated with more than 50% of fetal mortality.⁴³ If fetal death occurs, fetal heart rate monitoring usually is ominous at admission. On the other hand, five of eight women with a fetal death after an episode of domestic violence had a maternal injury severity score equal to zero.⁴⁷

Domestic and Sexual Abuse

Domestic and sexual abuse occurs at alarming frequency during pregnancy.⁴⁸ The incidence of domestic violence against pregnant women is 10% to 35%, depending on the population studied and the methods used to identify the abuse. The incidence and characteristics of abuse during pregnancy was examined in a stratified, prospective analysis of 691 African-American, Latino, and white pregnant women attending urban public prenatal clinics in Houston and Baltimore.⁴⁹ Physical or sexual abuse during pregnancy occurred in 17%. The abuse was recurrent in 60%, focused on the victim's head, was associated with drugs or alcohol, was highest among white women, and was associated with late prenatal care. In a larger study using the same screening tool on a similar population, the same authors found an incidence of abuse of 21% among pregnant teenagers and 15% among adults. Pregnancy complications, low birth weight, and substance abuse were more common in abused pregnant women than in nonabused pregnant women.⁵⁰

The reported incidence of sexual abuse also varies considerably. Among 329 Hispanic women, 32% reported sexual abuse by their male partner at least once in the previous year.⁵¹ On the other hand, Satin and associates⁵² report data from an urban university setting indicating that 1% to 3% of sexual assault victims are pregnant. They performed a retrospective review of sexual assault on pregnant victims in Dallas County, Texas, between 1983 and 1988. Of 5734 total victims, there was a total of 114 pregnant victims (2%, 0.55 pregnant assault victims per 1000 deliveries in Dallas County). The 114 pregnant assault victims were compared with matched nonpregnant sexual assault victims. The pregnant assault victim's obstetric outcomes were compared with the general obstetric population at Parkland Hospital. The typical victim was a black, parous gravida in her 20s, and at a mean gestational age of 15 weeks. The location of penetration (vulvar, 95%; oral, 27%; anal, 6%) and the detection of sperm were similar between pregnant and nonpregnant victims. Physical trauma was less common among pregnant victims (43% versus 63%; p = .004), especially genital trauma (5% versus 21%; p < .001) than among nonpregnant assault victims. Preterm delivery (16%) and low-birth-weight neonates (24%) were more common among pregnant assault victims than among the general obstetric population.

Unfortunately, obstetricians ascertain poorly the risk of domestic violence or depression.^53,54 Only 39% of obstetricians routinely screen their patients for domestic violence.⁵³ In an investigation of 41 injury-related maternal deaths in North Carolina,⁵⁴ obstetric care providers were not aware or suspicious of abuse in 67% of homicides committed by her intimate partner. The obstetric provider was aware of depression in only two of five suicide deaths. The obstetrician is responsible for screening all women. An introductory statement lets the patient know that your questions are universal: “I would like to ask you a few questions about your physical, sexual, and emotional trauma because we know that these are common and affect women's health.” Subsequently, four simple questions aid in the identification of women at risk: Has anyone close to you ever threatened to hurt you? Has anyone ever hit, kicked, choked, or hurt you physically? Has anyone, including your partner, ever forced you to have sex? Are you ever afraid of your partner⁵⁵?

Automobile Accidents (Blunt Trauma)

Blunt trauma is a common denominator to many injuries during pregnancy. Motor vehicle accidents are one source of blunt abdominal trauma and are a common source of injury during pregnancy. Crosby⁵⁶ reviewed 441 pregnant victims of automobile collisions; collisions were divided into the categories of minor or severe. When there was minor damage to the vehicle, only 3 of 233 victims had injury, and no placental separation occurred. When the damage to the vehicle was severe, 15 of, 208 pregnant victims (7.2%) died, and 25 of the 193 (13.5%) survivors had major injury. Fetal death after 12 weeks' gestation occurred in 14 of 176 (8%) mothers surviving a severe collision. Except for cases of maternal death, first-trimester loss could not be explained by trauma. The most common causes of fetal death after 12 weeks' gestation were maternal death, abruptio placentae, and maternal shock. Abruptio placentae occurred in 3.4% of victims involved in severe accidents.

Despite the recommendations of public health officials and obstetricians, the lay public and sometimes misinformed physicians raise the concern that restraints increase the likelihood of injury. Wolfe and coworkers⁵⁷ provide scientific support for the public health recommendations for women to use three-point restraints. Low birth weight (odds ratio = 1.9 [95% confidence intervals = 1.2, 2.9]) and delivery within 48 hours (odds ratio = 2.3 [95% confidence intervals = 1.1 4.8) were more common among 1349 unrestrained pregnant women (more than 20 weeks' gestation) than among 1243 restrained pregnant women. The lap belt should be placed over the pelvic bones and not over the uterus. The shoulder belt should be placed on the side of the uterus.

Air bags are a new addition to passenger protection in motor vehicle accidents. They save many lives when they deploy in severe accidents. Three-point restraints are 42% effective in preventing fatalities; the addition of a driver-side air bag provides a 12% increase in effectiveness.⁵⁸ Unfortunately, the occasional injuries that may be associated with their deployment have obscured the value of air bags to the public. Pregnant women are especially concerned about the impact to their pregnant uterus and its passenger. Overall, pregnant women are encouraged to use seat belts and not to deactivate the air bags during pregnancy.⁵⁸

Pearlman and Viano⁵⁸ performed a prospective cohort study that matched, by gestational age, 85 pregnant trauma victims with 85 pregnant women without trauma. Of the 85 trauma victims, motor vehicle accidents accounted for 60% (75% used three-point restraints), falls for 26%, and direct blow to the abdomen for 14%. Abruptio placentae, rupture of membranes, or both resulted in 8 of 85 (9.4%) delivering within 72 hours of the accident. Term (12%) mothers had moderate to severe injury, but there was only moderate correlation between the severity of the accident and pregnancy compromise; relatively minor trauma can cause pregnancy complications. Maternal—fetal transfusion (positive Kleihauer-Betke assay) occurred more frequently in trauma versus control patients (31% versus 8%; p < .05), especially if the placenta was placed anteriorly (47% versus 24%; p < .05). Of the 35 women who had contractions 2 to 5 minutes apart during the first 4 hours after admission, 5 (14.3%) had abruptio placentae.

Twenty percent to 30% of pregnant trauma patients show evidence of fetomaternal hemorrhage.^59,60,61 This complication can result in fetal anemia, chronic asphyxia, or Rh sensitization in Rh-negative women. Pregnant women with trauma should have a serum alpha-fetoprotein determination and a Kleihauer-Betke test to estimate the volume of fetal-to-maternal transfusion. Serial fetal testing and ultrasonographic examinations are necessary when large bleeds are apparent. In Rh-negative women, the standard 300-μg dose of RhoGAM covers a 30-mL fetal-to-maternal bleed. Only 0.1 mL of fetal blood is required for Rh sensitization.

Most labor and delivery units manage pregnant women who have experienced noncatastrophic trauma. Much time and money are spent on the workup and observation of these patients. In a series of 85 women admitted to a labor and delivery unit after noncatastrophic trauma,⁶² 13 (15%) delivered at less than 37 weeks. The clinical courses of the women delivering preterm were compared with those delivering at term. The Kleihauer-Betke stains, maternal vital signs, blood cell count, coagulation profile, and ultrasonographic examination results were normal in all 85 cases. There were no significant differences between the two groups regarding gestational age at the time of trauma, length of hospital stay, subjective reports of pain or tenderness, patterns of contractions, intervals between trauma and delivery, and Apgar scores.⁶²

Regardless of the source or severity of trauma, the mother and her fetus benefit most by knowledge, planning, and protocols that reflect the physiologic differences between pregnant and nonpregnant women and the special needs of the vulnerable fetus. After potentially catastrophic accident, a pregnant trauma victim should be observed in a hospital equipped to monitor the mother and the fetus. She should be transported on her side to avoid supine hypotension. A brief description of the severity of the accident and status of the patient should be obtained from the ambulance attendant. A brief physical examination should focus on central nervous system function (neck injury), vital signs, chest wall motion, and movement of extremities. At this point, diagnostic and monitoring studies should be performed. A large-bore intravenous catheter (16 gauge) is placed peripherally; if intra-abdominal trauma is suspected or if the patient is unconscious, a central venous pressure line should be placed. A complete blood count, urinalysis, serum electrolyte level, and blood for typing and cross-matching should be obtained. Tetanus toxoid (0.5 mL) should be given to all patients. Endotracheal intubation should be performed on all patients in respiratory failure or who are unconscious. A three-way Foley catheter should be placed in the bladder to monitor hourly urinary output and to identify hematuria. A nasogastric tube should be inserted to establish the integrity of the upper gastrointestinal tract and to empty the stomach of its contents. After 24 weeks' gestation, an external fetal monitor should be applied as soon as possible. During the placement of lines and monitors, a more complete history should be obtained and a physical examination performed.

The patient's vital signs should be monitored frequently, and a trained attendant should be with her at all times, including when she is taken for diagnostic x-ray studies. Young, healthy patients maintain their cardiovascular status up to a point at which shock develops quickly. X-ray and obstetric ultrasonographic studies are important. No ordinary, indicated x-ray examination should be avoided during pregnancy. Irradiation of the fetus poses little risk compared with the dangers of undiagnosed maternal trauma. Standard studies should focus on areas of injury. The obstetric ultrasonogram should confirm gestational age and fetal and placental position.

Fetal heart rate can act as a monitor of fetal and maternal well-being, adequacy of the gravida's circulating blood volume, and her compensatory α-adrenergic response. Fetal bradycardia (heart rate less than 120 beats/min), fetal tachycardia, or late decelerations may result from maternal hypovolemia, maternal hypoxia, abruptio placentae, or uterine rupture. In pregnancies of more than 25 weeks, cesarean section should be considered for fetal distress if the mother is stable. Fetal monitoring should continue for 24 hours because signs of fetal compromise can occur up to 24 hours after a major accident or injury.

In most trauma victims with multiple injuries or suspected intra-abdominal trauma, peritoneal lavage should be performed.^20,42,63 The procedure can be used in all trimesters. In the first trimester, the abdomen can be catheterized through the midline subumbilical percutaneous route. In obese patients, a tap through the umbilicus may be easier. In the second or third trimester, a semi-open technique is advised; the peritoneum is exposed under local anesthesia through a midline 3- to 4-cm incision. In either case, the abdominal cavity is entered with a peritoneal dialysis catheter. The catheter is advanced caudally, and if there is no return of fluid, 1000 mL of warm lactated Ringer's solution is infused during a 5-minute period. The tubing and bottle then are placed on the floor, and the fluid is allowed to siphon back to the bottle. The average return is approximately 750 mL. A positive lavage finding is suggested by free aspiration of blood (more than 10 mL), an erythrocyte count above 100,000/mm³, a leukocyte count higher than 175 cells/mm³, or the presence of gastrointestinal contents or an elevated amylase level (over 175 mg/dL) in the effluent.^20,42,63 A rapid negative test result can be achieved by the clinician's ability to read newsprint through the effluent. Even with a negative finding on lavage, significant injury can be present. The false-positive and false-negative rates are approximately 5%.^20,42,63 A negative lavage finding means that laparotomy need not be performed as long as bowel sounds are present and the abdomen remains soft and nontender.

The indications for exploratory laparotomy in trauma are a positive lavage result, free air under the diaphragm (before lavage), progressive abdominal distention with a declining hematocrit, or abdominal wall disruption or perforation. Intraoperative management of surgical trauma is specific to the type of injury involved. The reader is referred to surgical textbooks for specific techniques.

Stab and Gunshot Wounds

Stab and gunshot wounds are the most common major injuries to pregnant women other than those caused by automobile accidents. Violence against pregnant women often focuses on the pregnant uterus. Buchsbaum⁶⁴ reviewed 199 cases of gunshot wounds. Seventy percent of fetuses were injured, whereas only 28% of mothers had extrauterine injuries. The perinatal mortality rate was 64%, whereas only 3.2% of the mothers died. The extent of internal injury must be evaluated in all penetrating wounds to the torso, even if there is a clear entrance and exit wound. Most stab wounds that enter the thoracic or abdominal cavity require an exploratory operation. Penetration of the cavities can be documented radiographically. A catheter is placed deep in the wound and held in place with a purse-string suture. Rapid injection of water-soluble contrast material with subsequent x-ray films demonstrate entry. If entry has occurred, exploration is required. Gunshot wounds require more aggressive management because debris often is carried deep into the wound. Removal of debris and debridement of necrotic tissue are essential to reduce infection. Anterior and posterior radiographs of the torso demonstrate the position of radiopaque debris (e.g., bullet fragments).

If the mother's condition is stable, the fetus must be evaluated before exploration. After 24 weeks' gestation, an electronic fetal monitor is used to evaluate fetal health.⁶⁰ Fetal heart rate accelerations (above 10 beats/min for more than 15 seconds at 24 to 31 weeks, above 15 beats/min for more than 15 seconds after 31 weeks) are reassuring for the fetus. A gunshot wound to the maternal torso is an indication for amniocentesis under ultrasonographic guidance. Bloody amniotic fluid suggests fetal injury. If the fetus is at 34 weeks' gestation or greater, a rapid test (OD 650 or other rapid amniotic fluid test) of fetal lung maturity may be helpful in decision analysis.

At the time of exploratory laparotomy, the uterus is examined for penetrating injury. If a penetrating injury is present, if the fetus is older than 25 weeks, and if there is evidence of fetal compromise (e.g., bloody amniotic fluid, abnormal fetal testing), cesarean delivery is indicated. A neonatologist should attend the preterm delivery. If the fetal lungs show maturity, cesarean delivery is indicated for any penetrating injury to the uterus regardless of the results of fetal monitoring. However, if the amniotic fluid is clear but shows immature lungs and the monitor indicates fetal well-being, the uterus may be repaired and the fetus left alone. Labor may ensue after recovery from anesthesia. In the absence of vaginal bleeding or fetal distress, labor may be suppressed with magnesium sulfate.

Penetrating uterine injury at less than 25 weeks should be treated conservatively. Delivery would mean almost certain neonatal death from prematurity, and fetal fractures and stab or bullet wounds may heal in utero. The only reason for hysterotomy is maternal hemorrhage or fetal death in association with a uterine laceration that would preclude labor (e.g., a large fundal laceration). In general, vaginal delivery is preferable to hysterotomy in cases of fetal death, even if delivery will occur soon after exploration.

Burns

Pregnant women rarely are burned seriously, but when they are, they have unique medical problems.^65,66 The mother and fetus are at great risk for fluid loss, hypoxemia, and sepsis. In general, pregnancy does not alter outcome; maternal survival is accompanied most often by fetal survival. The overall fetal and neonatal mortality rate is over 50% when the mother is burned over more than 60% of her body.

Complications of burns relate to the total surface area burned and the depth of the burn. A general estimate of the body surface area involved by a burn is determined by the rule of nines: head and neck (9%), upper extremities (9%) each, anterior torso (18%, first 20 weeks; 23%, second 20 weeks), posterior torso (18%), lower extremities (18% each), and genitalia (1%). Another method is to equate the number of palmar surfaces the burn entails, each palmar surface being equal to 1.25% of body surface. During late pregnancy, 5% is added if the anterior abdomen is involved. The depth of the injury is estimated by appearance and sensation. Partial-thickness (intradermal) injury appears red or pink with blister formation. Full-thickness injury may be charred or marble gray in color, dry, and anesthetic. Partial-thickness burns also may be anesthetic because of neuropraxias of skin nerve endings in the burn area. Thus, pain response to stimulation is valuable only to indicate a partial-thickness burn.

Fluid replacement, respiratory support, and initial wound care are the emergency management goals in pregnant burn victims. The loss of fluid through the denuded surface can be massive, and the amount often is underestimated in pregnant patients. On arrival at the hospital and after the vital signs of the mother and fetus (monitor) are taken, a large-bore (18-gauge) intravenous line is started. In cases in which the burn covers more than 20% of the surface area, a central venous or Swan-Ganz catheter provides a better guide to fluid replacement. Lactated Ringer's solution is started at 200 mL/hr until the fluid replacement volume is calculated. A nasogastric tube should be inserted for burns involving more than 20% of body surface area. Hospital admission is recommended for smoke inhalation, electrical burns, burns of both hands or both feet, partial-thickness burns of over 10% of the surface area, or full-thickness burns on more than 2% of the surface area.

When the burn surface is over 15% of the total body surface, initial fluid resuscitation is 1000 mL/hr of lactated Ringer solution and titrated to maintain a urine output of 0.5 mL/kg/hr. The free-water requirement (500 mL) is supplied with 5% dextrose in water. Fifty percent of the replacement fluid is given in the first 8 hours and the remainder during the next 16 hours. In the second 24 hours, colloids (albumin) are given to maintain the serum albumin higher than 3 g/100 mL.

Fluid replacement is monitored by clinical and laboratory means. Systolic blood pressure should be above 110 mmHg, maternal heart rate less than 120 beats/min, temperature less than 38°C, and respiratory rate 12 to 24 breaths/min. Central venous pressure should be approximately 10 cm H₂0, and urine output should be more than 0.5 mL/kg/hr. The initial laboratory workup should include a complete blood count and determination of blood levels of electrolytes, glucose, albumin, urea nitrogen, and serum creatinine. These parameters should be monitored on a serial basis (every 4 to 8 hours).

Smoke inhalation is a major cause of morbidity and mortality in burn victims. In pregnancy, the fetus is at special risk because of its relatively hypoxic state (normal umbilical vein Pao₂ = 27 mmHg). The pathophysiologic course of inhalation injury relates to impaired maternal ventilation (upper airway obstruction from edema), increased diffusion distance (interstitial alveolar edema), and acute functional anemia from carbon monoxide poisoning. Carbon monoxide binds more efficiently to hemoglobin than does oxygen. In addition to displacing oxygen, carbon monoxide impairs the release of oxygen from oxyhemoglobin. Little carbon monoxide is needed to cause serious hypoxia. One part carbon monoxide per 1500 parts air can result in blood concentrations of carboxyhemoglobin of 5% to 10%. Car exhaust is 5% to 7% carbon monoxide. Carboxyhemoglobin values less than 15% usually are well tolerated, whereas values over 30% cause severe maternal syncope and fetal death.

Inhalation injury should be suspected among patients who have a history of closed-space fire, facial injury, carbonaceous material in the oropharynx, or respiratory symptoms. Interstitial edema on chest x-ray film, a carboxyhemoglobin level above 10%, or abnormal arterial blood gas levels also aid the diagnosis of inhalation injury. Initial management of any burn patient should include an arterial blood sample for gases and carboxyhemoglobin, as well as a chest radiograph. Patients should be placed on 100% oxygen by mask for at least 3 hours or until the carboxyhemoglobin level is known. They should receive vigorous chest physiotherapy. Intubation and mechanical ventilation should be used early in the presence of upper airway obstruction or oxygenation failure.

Sepsis is another major risk for the fetus and mother. Initial wound care can be instrumental in the prevention of these complications. On admission, the wound is cleaned with bland soap and water, and all dirt and loose devitalized tissue are removed. Blisters should be left intact if they are smaller than 5 cm in diameter. When burns involve the scalp, axilla, or pubic area, the hair should be clipped short until an adequate margin of unburned skin is obtained. After cleaning and debridement, a topical agent is applied with a bulky dressing. Silver sulfadiazine cream is used most commonly, but the consideration in pregnant patients is that this drug can be absorbed: the sulfa derivative crosses the placenta and displaces bilirubin. Should delivery ensue, hyperbilirubinemia is a risk for the neonate. Silver nitrate (0.5%) also is used, but this agent requires continuous soaking (every 2 hours) and a bulky dressing. Tetanus toxoid (0.5 mL) should be given to all burn victims.

After the initial management of a severely burned patient, her care requires a team approach with the obstetrician acting as a consultant. Pregnant women with severe burns are best cared for in centers geared both to managing severe burns and to the possibilities of early delivery. The major long-term problems are healing, sepsis prevention, scar complications, nutritional support, and rehabilitation. Contrary to poplar belief, major burns (full thickness to over 50% of the abdominal wall) in childhood burns does not alter the accommodation of the abdominal wall to expansion in the third trimester.⁶⁷ The various methods and problems of long-term care are beyond the scope of this chapter; the reader is referred to a textbook on the management of burn victims.

Intracranial Hemorrhage

Intracranial hemorrhage is a neurovascular emergency requiring surgery in many pregnant women. Although the incidence of congenital malformations of cerebral vasculature, arteriovenous malformations (AVMs), and aneurysms is uncommon (0.01% to 0.05% of all pregnancies), traumatic head injury and intracranial hemorrhage associated with eclampsia significantly increase the incidence of neurosurgery in pregnancy. This section focuses on vascular malformation because of its predictable clinical course and the negative influence of pregnancy on the disease.

A recent review of 154 pregnant patients with intracranial hemorrhage found aneurysms to be the cause in 118 (77%) patients and AVMs in 36 (23%). The distribution and location of such lesions appear to be similar between pregnant and nonpregnant women.^68,69 Patients with aneurysms tend to be older than those with AVMs. The risk of hemorrhage in both lesions is greatest in the second and third trimester. In patients with a previously asymptomatic AVM, the risk of hemorrhage during pregnancy is 3.5%—not significantly different from the annual bleeding rate in the nongravid patient with an unruptured AVM. Unfortunately, hemorrhage from an AVM results in a higher maternal mortality rate (28%) than it does in the nongravid population (10%).^68,69,70,71 The maternal mortality rate in intracranial aneurysmal bleeding is 35%.^68,69,70,71 Despite the concern and focus on the intravascular pressure changes associated with labor and delivery, the association between these dramatic but short-lived hemodynamic changes and hemorrhage is not clear: only 3% of intracranial bleeds occur during labor and delivery.⁶⁸ Recurrent bleeding during the remainder of pregnancy from an untreated aneurysm or AVM is estimated to occur in 35% to 50%, with a maternal mortality rate of 50% to 60%.^69,72 In a controlled study of 55 women whose aneurysm was clipped during the pregnancy compared with 51 women whose aneurysm was not treated during pregnancy, the maternal mortality (11% versus 63%) and fetal mortality (5% versus 27%) was significantly lower in the treated group than in the untreated group.⁶⁸ In the same study, 13 patients with an AVM were treated and 23 patients were not treated surgically. The differences in maternal mortality (23% versus 32%) and fetal mortality (0% versus 23%) were not statistically different. Despite the observation concerning the treatment of aneurysms and AVM, the decision whether and when to operate on an aneurysm or AVM in the gravid patient should be the same as for the nongravid patient: it is a neurosurgical decision, not an obstetric one.

The signs and symptoms of a subarachnoid bleed include the following: acute onset of a severe headache, the patient's description as the “worst headache of my life,” nausea, vomiting, stiff neck, photophobia, seizures, and a decreasing level of consciousness. Focal neurologic deficits may not be apparent immediately; 40% of patients with aneurysmal hemorrhage have no localizing findings. Focal deficits that develop 3 to 10 days later result from cerebral ischemia from reactive vasospasm.

Once an intracranial hemorrhage is suspected clinically, cranial CT scanning without contrast is recommended. The contrast dye is too dense on a CT scan and can be confused with blood, thus leading to misdiagnosis. Contrast agent may be useful after a noncontrasted CT scan. Magnetic resonance imaging is not as helpful because fresh blood in a hemorrhage is not easily contrasted. Angiography is the standard method to locate and characterize an intracranial hemorrhage. In addition, it allows measurement of vascular spasm—an important consideration in surgical planning.

Recurrent bleeding and vasospasm are the two largest causes of death and disability after the original intracranial bleed.^69,72 Early (less than 24 hours) rebleeding rates are highest in aneurysmal hemorrhage (5% to 10%) and approach 50% in the remainder of the pregnancy. Most data on AVM rupture suggest a lower early and late rebleeding rate: 1% and 4%, respectively. Vasospasm occurs in one third of patients with aneurysmal hemorrhage but rarely after AVM rupture. Vasospasm is manifested by clinical deterioration and decreasing consciousness or focal neurologic deficit starting 3 to 6 days after the initial hemorrhage and is not attributed to other causes such as rebleeding, hydrocephalus, or metabolic disturbance. Vasospasm reaches a peak at 10 to 12 days and can last for 2 to 3 weeks. Medical and surgical treatment is required.

Current neurosurgical management emphasizes early surgical treatment of aneurysmal rupture among patients with Hunt and Hess grades I to III and accessible lesion. Immediate surgical intervention in Hunt and Hess hemorrhage grades IV and V or in difficult locations (posterior fossa) is more controversial. Often, surgery is delayed 3 to 4 weeks until neurologic improvement occurs.

Surgical treatment of accessible and less severe lesions is coupled with aggressive management of vasospasm: induced hypervolemia and the calcium-channel blockers, nimodipine, and nicardipine. The danger of these treatments in pregnant women is the induction of pathologic hypervolemia and pulmonary edema, with subsequent maternal and fetal hypoxia. Calcium-channel blockers have had limited use in pregnancy; nifedipine has been used primarily in the treatment of hypertensive disease. Although some evidence links nifedipine to fetal acidosis in ovine models, the potential benefit of calcium-channel blockers in preventing and controlling death and disability from vasospasm outweighs the theoretic risk. Nimo

Prolonged malnutrition from gastrointestinal disease can lead to intrauterine growth retardation. Most patients with an acute abdomen are not permitted oral intake, at least for the acute phase, and intravenous dextrose and multivitamins are indicated. If a patient has had insufficient intake (less than 1500 kcal/24 hours) for 72 hours, parenteral nutrition should be started. If support will be needed for less than 10 days, peripheral parenteral nutrition is indicated. An appropriate nutrient solution is the simultaneous use of 4.25% amino acid and 5% dextrose (1500 mL) with a fat emulsion (1500 mL of lntralipid) through a Y-connector. This provides the patient with 3000 mL of water, 64 g of protein, and 1950 calories in 24 hours. Electrolytes, trace elements, and vitamins can be added to the protein solution. If support will be needed for more than 10 days or if a patient's condition is severe (e.g., recurrent Crohn's disease), total parenteral nutrition through a central venous catheter is needed.

The specific problems of parenteral nutrition during pregnancy are the increased metabolic needs (increased to 40 kcal/kg/day and an increase in protein by 30 to 50 g/day). Hyperglycemia can be a greater risk because pregnancy normally is a diabetogenic state. Mean blood glucose levels consistently higher than 105 mg/dL should be treated with insulin. Pregnancy is a hyperlipidemic state, and there has been some theoretic caution about the use of lipid solutions (e.g., pancreatitis, preterm labor). Hyperalimentation, however, is a replacement rather than a supplementation, and this caution may not be applicable. During hyperalimentation, the fetus is evaluated for growth and well-being; nonstress tests are performed twice weekly, and serial ultrasonography for fetal growth should be obtained.

Preterm Labor

Preterm labor may be the cause of abdominal pain or a consequence of nonobstetric disease, trauma, or infection. The diagnosis and treatment of preterm labor among pregnant patients perioperatively are difficult. There are clear risks for preterm delivery; however, the intrauterine environment may not be ideal, and the treatment of preterm labor may exacerbate the effects of intra-abdominal disease (e.g., appendicitis). Before making a definitive diagnosis, the many obstetric conditions that pose greater fetal risk (e.g., chorioamnionitis, abruptio placentae) must be considered. These often mimic acute nonobstetric conditions. Tocolysis in the presence of these diseases is not indicated. The cardiovascular effects of β-adrenergic drugs (e.g., terbutaline, ritodrine) can exacerbate those of sepsis (e.g., vasodilation, shock). For example, intravenous ritodrine poses a 5% to 10% risk of serious cardiopulmonary complications such as pulmonary edema and significant subendocardial ischemia. Other tocolytic agents that involve calcium ion physiology (e.g., magnesium sulfate, calcium-channel blockers) have less significant but still disturbing cardiovascular effects in the presence of sepsis. Prostaglandin inhibitors (e.g., indomethacin) are contraindicated for tocolysis because they mask the clinical signs of infection.

Keeping these problems in mind, the following management of preterm labor in acute intra-abdominal conditions is suggested. First, uterine irritability is not considered to be labor until there is evidence of cervical effacement or dilatation. Prophylactic tocolysis is not indicated. Any decision to begin tocolysis requires demonstration of the following: cervical change, assurance that membranes have not ruptured through sterile examination using Nitrazine and fern tests, and the absence of the usual precautions against the use of tocolytics. Second, the presence of intrauterine disease should be determined by a thorough ultrasonographic examination and, perhaps, amniocentesis. Amniocentesis under ultrasonographic guidance has many advantages. Bacteria or sheets of leukocytes on a high-power microscopic examination of an unspun specimen suggest amniotic fluid infection. Fetal lung maturity studies can be performed rapidly in most institutions. Keep in mind the risks of third-trimester amniocentesis: fetal trauma (1%), ruptured membranes (1% to 2%), and preterm labor (1% to 2%). If peritonitis is present, amniotic fluid contamination is a special concern. In the presence of peritonitis, an exploratory laparotomy should not be delayed by amniocentesis. Localized peritoneal contamination over the amniocentesis site cannot always he predicted. Amniocentesis should be performed away from points of tenderness. An uncomfortable but acceptable alternative is to perform a suprapubic tap through a full bladder under ultrasonographic guidance. This approach avoids the peritoneum.

Once the diagnostic test or surgery has been performed and the mother's cardiovascular status is stable, the decision whether to inhibit labor can be made. Magnesium sulfate is the therapy of choice. The dose is 4 g in 250 mL of normal saline given intravenously during a period of 15 minutes, followed by 2 g/hr by electronically monitored intravenous drip. The fetus should be continuously monitored throughout diagnosis and management. Fetal distress should be aggressively diagnosed and treated. Magnesium sulfate should be continued for 12 to 24 hours after the operation. At that point, patients are weaned off the medication.

Iatrogenic Infertility

First-trimester complications in an undiagnosed pregnancy bring many young women under the care of general surgeons unfamiliar with obstetric problems. Conservative management should be practiced. Oophorectomy is rarely indicated in pregnancy unless the tumor is large (more than 10 cm), solid, or exhibits signs of malignancy. Unless significant hemorrhage has occurred, oophorectomy should not be performed for a bleeding corpus luteum cyst. Incidental resection of a corpus luteum cyst at the time of laparotomy for other diseases should be discouraged because removal can cause spontaneous abortion (see later). Ovarian biopsy or cystectomy has a 35% incidence of iatrogenic postoperative adhesions.⁷⁷ The incidence of subsequent infertility is hard to determine, but it may be as high as 5% to 10%. When ovarian surgery is indicated, careful surgical technique is critical to reduce the risk of adhesions. These techniques include the use of fine (4-0 to 6-0) nonreactive absorbable sutures, atraumatic tissue handling, excellent hemostasis, and avoidance of tissue dehydration.

The use of adjunctive procedures to decrease adhesions has been suggested. The most popular are the instillation of 50 to 100 mL of 32% dextran and administration of systemic steroids and antihistamines after surgery. Theoretically, these procedures work by inhibiting contact adhesions and the inflammatory response. Unfortunately, well-designed prospective studies are needed to define their efficacy in pregnancy. In addition, the risks to the fetus presented by the use of dextran, steroids, and antihistamines are not well defined.

Fetal Assessment

Many causes of acute abdomen are associated with uteroplacental insufficiency (e.g., trauma, infection). Unless the condition is chronic (e.g., pancreatitis), the need for fetal monitoring is episodic. At any gestational age, documentation of a live fetus by fetal heart auscultation at regular intervals is appropriate. Continuous fetal monitoring for fetal heart rate changes is appropriate only if an obstetrician is willing to act on the information (e.g., fetus more than 25 weeks' gestation or with no lethal anomalies). In the recovery period, as a patient's condition becomes stable and consciousness is regained, continuous fetal assessment can be changed to a combination of recording 10 fetal movements every 12 hours and nonstress tests twice a week. This scheme can be used until the patient has returned to normal but not necessarily for the duration of the pregnancy.

Similarly, uterine contraction monitoring is appropriate only if the documentation of contractions would change management. In the acute phase, continuous uterine contraction monitoring is appropriate after 20 to 22 weeks' gestation with a normally formed fetus. Serial cervical examinations are essential for the diagnosis of preterm labor. Tocolytic therapy should be withheld until cervical change has taken place. If there is no uterine irritability, the patient's condition has stabilized, and she can recognize and report fetal movement and contractions, continuous uterine contraction monitoring can be replaced by maternal perception and patient education about the signs of preterm labor.

Surgical disease in pregnancy presents dilemmas in diagnosis, techniques, and postoperative management. These dilemmas result from normal changes in maternal anatomy and physiology and an appropriate concern for fetal well-being. Cooperation between a maternal—fetal medicine specialist, surgeon, and obstetrician maximizes the benefits of diagnosis and therapy for the mother and the fetus.