Chapter 25: AMNIOCENTESIS

Amniocentesis is the oldest invasive procedure for prenatal diagnosis. It has been in use for more than 100 years, being first employed in the 19th century for the treatment of polyhydramnios.^{1,2, and 3} Subsequently, amniocentesis was used to perform procedures such as amniography⁴ and elective terminations of pregnancy,^5,6 and it gained popularity in the last 5 decades as a diagnostic tool in the management of pregnancies complicated by isoimmunization⁷ as well as in cases of genetic defects. Indeed, Fuchs and Riis⁸ reported in 1956 prenatal gender determination by analysis of the sex chromatin of cells obtained by amniocentesis. Human cytogenetics emerged as a field in 1966 when Steele and Breg⁹ reported the successful determination of a human karyotype from cultured amniotic fluid cells. A year later, the first prenatal diagnosis of a chromosomal abnormality was reported: a balanced translocation.¹⁰ Trisomy 21 was initially diagnosed in 1968 by Valenti et al,¹¹ and the first congenital metabolic disorder diagnosed by amniotic fluid analysis was adrenogenital syndrome in 1965.¹²

Table 25-1 presents the most frequent indications for amniocentesis. The most common indication is prenatal diagnosis. Evaluation of fetal lung maturity has been the leading indication for the procedure in the third trimester. The role of amniotic fluid analysis in the diagnosis of intrauterine infection has increased over the last decade and is discussed in detail later in this chapter. Regardless of the indication for amniocentesis, all patients must undergo detailed counseling about the objectives, procedure, and potential complications of amniocentesis. An informed consent should be signed before the procedure.

Before the amniocentesis, an ultrasound examination is mandatory to determine fetal viability, gestational age, number of fetuses, placental location, adequacy of amniotic fluid volume, and the presence of any abnormality that might impact the performance of the procedure (ie, uterine leiomyoma, fetal anomaly, etc).

Gestational Age

Genetic amniocentesis is usually performed after 15 weeks (most commonly between 16 and 18 weeks of gestation). Before the introduction of ultrasonographic guidance, early studies of amniocentesis indicated a lower success rate in retrieving amniotic fluid when the procedure was performed at gestational ages equal to or less than 15 weeks.^13,14 The maternal risks associated with late second trimester termination of pregnancy and the psychologic trauma of terminating a pregnancy that is both visible and felt led to the search for alternative procedures for earlier prenatal diagnosis, namely chorionic villus sampling (CVS) and early amniocentesis.^{15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34, and 35} Early amniocentesis is discussed in detail in the next section. For a discussion of CVS, please refer to Chapter 24.

Needle Selection

Amniocentesis is performed with a regular spinal needle. It has been recommended that the needle bore should be between 20 and 22 gauge. Larger-bore needles have been associated with an increased incidence of fetal loss.^13,15 Smaller-bore needles prolong the period of time required to obtain amniotic fluid and are more difficult to guide if intraoperative manipulations are required. The standard length of a spinal needle is 8.89 cm (excluding the hub), but longer needles are required for procedures in obese patients. Currently, commercially available needles are optimized for sonographic visualization. Acoustic visualization can be improved by roughening the needle, and Teflon coating has been added to decrease needle friction. Needles with side orifices are also commercially available. This design allows fluid to flow not only through the needle tip but also through the side orifices. This may have some advantages in cases of severe oligohydramnios. Although in vitro studies have shown that the flow rate during aspiration doubles with 22-gauge needles with side holes when compared with standard spinal needles of the same caliber, the use of these needle types is strictly dependent on personal preference until an adequate clinical trial is performed.

Ultrasound

Ultrasound has become an integral part of the amniocentesis procedure. The term sonographically guided technique refers to the use of ultrasound for selecting the site of needle insertion before amniocentesis; then, the needle is inserted blindly into the amniotic cavity. In contrast, the term sonographically monitored technique describes the continuous use of ultrasound throughout the procedure so that the insertion of the needle and the movement of the fetus are under constant surveillance.³⁶ The sonographically monitored technique has been shown to reduce the frequency of bloody and dry taps (and multiple needle insertions) in comparison with the sonographically guided technique.³⁶ The monitored technique also adds to the ease and expedience of the procedure, improves the patient's understanding of amniocentesis, and allows the operator to correct for potential difficulties during the procedure, such as tenting of the membranes or uterine contractions.

Operative Technique

Figure 25-1 shows the instruments used to perform an amniocentesis. After the ultrasound examination, asepsis of the skin is performed with wide boundaries (Figure 25-2), and then the field is draped. Needle insertion is performed under sonographic visualization. Several techniques have been described for this purpose, and we have described a technique that continues to be employed in our institution.^36,37 A sterile coupling agent is applied to the skin (Figure 25-3), and a nonsterile coupling agent is applied to the surface of the transducer, which is placed into a sterile glove or a sterile plastic bag (Figures 25-4, and 25-5). A site for needle insertion is selected, while trying to avoid a localized contraction, a uterine leiomyoma, the umbilicus, the placenta or the umbilical cord. Once the insertion site has been selected, a finger is placed under the transducer (Figure 25-6). Decoupling of the transducer from the skin's surface produces a shadow that allows the identification of the needle path. Under direct sonographic visualization, the needle is inserted along the side of the transducer, and the tip, which appears as a bright echo, is continuously monitored throughout the procedure (Figure 25-7). The stylet is removed, and a syringe is attached to the needle to aspirate the amniotic fluid (Figure 25-8). Alternatively, an extension tube is attached to the hub of the needle and connected to the syringe (Figure 25-9). The purpose of the plastic tube is to allow the needle to float freely in the amniotic cavity, thereby decreasing the likelihood of fetal injury. This method also prevents operator movements (ie, sneezing) from being transmitted to the needle. The first 0.5 cc of amniotic fluid is discarded to decrease the likelihood of contamination of the fluid with maternal cells. After obtaining the fluid, the stylet is replaced and the needle is removed. Fetal cardiac activity should always be documented with ultrasound after the procedure.

Needle-guiding devices have been introduced as another technical modification for use in invasive prenatal diagnosis. Some investigators have reported reduced risk of amniocentesis with devices that affix to the ultra-sound transducer and mechanically guide the needle used to aspirate amniotic fluid. Since this modification can limit the operator's ability to maneuver the needle, the technique has been further modified to incorporate an articulated needle guide to combine the advantages of the needle-guided procedure and the free-hand technique.

At the conclusion of the procedure, either the syringe or the tubes in which the amniotic fluid will be transported to the laboratory are labeled with the name of the patient. The patient is asked to verify the identification of her own sample. An amniocentesis report should include the ultrasound findings, number of needle insertions, color, and volume of the amniotic fluid retrieved, whether or not the amniocentesis was transplacental, and any other unusual findings or events occurring during the procedure. The patient is instructed to report any signs of infection or vaginal leakage of fluid. There is no evidence to support restriction of normal activity after an unevent-ful procedure. If the patient is Rh-negative and not sensitized, Rh immunoglobulin is administered.

An anterior placenta does not contraindicate the procedure, but if a transplacental puncture is necessary, preference is given to the thinnest portion of the placenta. The potential role of transplacental needle passage in procedure-related pregnancy loss in early and second trimester amniocenteses has been assessed. A study of 380 cases of amniocenteses before 14.9 weeks at the University of Tennessee, found 147 cases of transplacental needle passage (38.7%).³⁸ The frequency of bloody fluid was higher in cases of transplacental (n = 6) versus nontransplacental (n = 1) needle passage (P < .05). The rate of pregnancy loss, however, did not differ between cases of transplacental (3.6%) and nontransplacental (3.6%) needle passage. These results were similar to those obtained in second trimester amniocentesis. An investigation of 1000 consecutive patients referred for genetic amniocentesis showed a nonsignificant difference in pregnancy loss between 306 cases with transplacental needle passage (1.96%) and 694 cases without transplacental needle passage (1%, P = .23).³⁹ Similarly, a retrospective review of 2083 second trimester amniocenteses showed no increased risk of pregnancy loss for 476 patients exposed to transplacental needle passage.⁴⁰ None of the studies had adequate power to rule out a difference between groups, and the investigators reported that, when transplacental needle passage was unavoidable, the thinnest accessible portion of placenta was traversed. Amniotic fluid contamination with maternal cells is regularly observed in pregnancies with an anterior placenta, whereas it is rare in cases with or posterior placenta. The maternal cells are therefore thought to be artificially introduced into the amniotic fluid sample during the transplacental puncture.

Local anesthesia is not routinely employed at our institution. Although some patients may benefit from its use, a limitation of local anesthesia is that it requires an additional needle puncture. Moreover, Van Schoubroeck and Ver-haeghe⁴¹ conducted a randomized clinical trial including 220 women who underwent genetic amniocentesis between 14 and 20 weeks of gestation. Patients were randomized to receive local anesthesia (lido-caine 1%) or no treatment. The authors demonstrated that there was no significant difference in the mean pain score measured by visual analog scale (VAS) between patients that received local anesthesia and those who did not (1.4 ± 1.5 vs 1.3 ± 1.4, respectively). Interestingly, 97% of patients described the procedure as not painful or tolerable, 79% had anticipated the procedure to be more painful, 59% reported the amniocentesis to have a comparable discomfort as venous blood sampling, and 98% of women stated they would undergo an amniocentesis again if indicated. These findings support the view that pain associated with amniocentesis is usually mild. Harris et al⁴² reported that 62% of women undergoing genetic amniocentesis did not have pain or the pain was described as mild, while 31% described the pain as discomforting, and only 7% described it as distressing or horrible. The mean intensity of pain was 1.6 ± 1.3 (on a scale of 0 to 7). Maternal weight, parity, previous surgery, fibroid tumors, depth of needle insertion, and presence or absence of an accompanying person were not associated with pain intensity. In contrast, pain during amniocentesis was associated with increased maternal anxiety, a history of menstrual cramps, a previous amniocentesis, and insertion of the needle in the lower uterus. Indeed, the degree of pain may depend on location of needle insertion. In a recent study,⁴³ women undergoing mid-trimester amniocentesis were asked to complete a VAS after the amniocentesis. Location of needle insertion was defined as "upper third" or "middle third" based on the distance from uterine fundus to the pubic symphysis. The mean VAS score was 2.7 ± 2.1, and perception of pain was significantly less in patients with needle insertion in the upper third as compared with that of those in the middle third (VAS 2.2 vs 3.9, respectively; P = .002). Previous amniocentesis, transplacental amniocentesis, previous abdominal surgery, and operator's experience were not associated with pain intensity.⁴³

Volume of Amniotic Fluid

The volume of amniotic fluid drawn at the time of genetic amniocentesis varies from 8 to 45 cc. Most centers retrieve between 20 and 25 cc. This represents approximately 10% of the mean volume of amniotic fluid at 16 weeks of gestation. The effect of amniotic fluid volume on pregnancy loss has been the subject of contradictory reports. Whereas the American Collaborative Study concluded that the volume removed was unrelated to fetal loss,¹⁵ the Canadian Trial suggested an increased prevalence of neonatal complications with volumes greater than 16 cc.¹³ The volume of amniotic fluid retrieved in centers performing early amniocenteses (before 15 weeks) has ranged between 8 and 25 cc. The recommended amount of fluid to be withdrawn is 1 mL per week of gestation. The first 0.5 to 1 mL of fluid aspirated should be discarded to decrease the risk of maternal cell contamination. The time required to replace the volume of fluid aspirated has not been established. Although a 3-hour period has been cited as an average replacement time for mid-trimester amniocentesis (16 to 18 weeks), there are no adequate data to support this view.

Membrane Tenting

The term membrane tenting refers to the separation of the chorioamniotic membrane from the anterior uterine wall during needle insertion. The membranes are carried forward by the needle, but without penetrating them. The diagnosis is made when the tip of the needle is visualized within a clearly defined pocket and no fluid is obtained. Tenting of the membranes is a frequent cause of amniocentesis failure and multiple needle insertions that can be overcome by twisting or redirecting the needle. If these maneuvers fail, another approach is to advance the needle into the posterior uterine wall, physically displacing the obstructing membrane down the shaft and away from the tip. Occasionally, patients will be referred to a different center when amniocentesis has failed and significant chorioamniotic separation has occurred. Amniocentesis can be deferred until reattachment is documented (1 or 2 weeks) or a transplacental needle insertion can be attempted, because membrane tenting is unlikely to occur at this site since the chorioamniotic membranes are more adherent on the surface of the chorionic plate. Dombrowski et al⁴⁴ reported a modified stylet technique to overcome failed aspiration of amniotic fluid in cases complicated by tenting of amniotic membranes. The technique of advancing a stylet from a 12.7-cm spinal needle through the original 8.9-cm 22-gauge spinal needle yielded successful aspiration of fluid in 21 of 22 cases of tenting of amniotic membranes. This technique can be used to avoid multiple needle insertions or postponement of the procedure when tenting of membranes persists despite rotating or advancing the needle. Johnson et al⁴⁵ reported an association between amniotic membrane tenting and abnormal karyo-type, specifically trisomy 21. Tenting was significantly more frequent among trisomy 21 fetuses (29%) compared to only 9% of other chromosomally abnormal pregnancies, and only 8% when the karyotype was normal.

Multiple Needle Insertions

The American Collaborative Study⁴⁶ of mid-trimester amniocentesis reported an increased frequency of spontaneous abortion and stillbirth after multiple needle insertions (more than 2). Similarly, the Cana-dian Trial¹³ found an association between multiple needle insertions (more than 2) and total fetal losses. Other studies examining this relationship, however, have not confirmed these findings. Since the introduction of sonographic monitoring of amniocentesis, the frequency of dry taps and multiple needle insertions has decreased significantly. For example, only 1.9% of all patients in a trial conducted by Tabor et al⁴⁷ required more than 1 needle insertion, and no patient required more than 2 needle insertions. These investigators could not demonstrate an association between the number of taps and fetal loss. Another study⁴⁸ including amniocenteses performed only in the third trimester reported that 9% of patients required more than 1 needle insertion (1.4% required 3 or 4 attempts). The need for multiple attempts was not associated to the level of training of the operator.

Bloody Taps

Bloody amniotic fluid can result from contamination with maternal blood, fetal blood, or both. The prevalence of this complication has also decreased with the use of the sonographically monitored technique. In third trimester amniocenteses, Gordon et al⁴⁸ found that the incidence of bloody amniotic fluid was higher if more than 1 needle stick was necessary (30% vs 8%, P = .001) or if the needle traversed the placenta (40% vs 3.6%, P = .001). Giorlandino et al⁴⁹ reported in repeated amniocenteses performed on 20 women 2 weeks after the first procedure (due to contamination of the first cell cultures) that blood-stained amniotic fluid could be observed in 75% of women. The authors found a significantly higher red blood cell count and hemoglobin concentration in the amniotic fluid of women who had previously undergone a transplacental amniocentesis when compared to those who did not have a transplacental procedure, and speculated that bleeding may occur after withdrawing the needle, leaving this contamination unrecognized in most cases. Yet it has been observed that, even if contamination with red blood cells occurs during amniocentesis, centrifugation usually prevents hemolysis and hemoglobin release.

The American Collaborative Study⁴⁶ did not demonstrate an increased risk of fetal loss with bloody taps, whereas most other studies did not comment on this issue (ie, the Canadian and British Collabora-tive Studies).^13,50 Ron et al⁵¹ specifically addressed the clinical significance of blood-contaminated amniotic fluid. They studied 706 women undergoing mid-trimester amniocentesis. The first 2 mL of amniotic fluid was examined for the presence of blood from maternal or fetal origin. The prevalence of bloody taps was 25.5% (180 of 706). Maternal contamination occurred in 84.4% (152 of 180) of cases, while fetal contamination was less frequent (15.6%; 28 of 180). A mixture of fetal and maternal blood was present in 8 cases. The incidence of spontaneous abortion was significantly greater in patients with bloody taps than in patients with clear taps (maternal blood contamination: 6.6%; fetal blood contamination: 14.3%; and clear taps: 1.7%). The incidence of pregnancy loss when fetal blood contamination occurred was twice as that observed when maternal blood contamination was documented. This difference, however, did not reach statistical significance. The investigators did not provide separate risk figures for microscopic versus gross bloody taps.

Fetomaternal Transfusion

Fetomaternal transfusion can be detected by performing a Kleihauer-Betke test after the procedure or by determinations of maternal serum α-fetoprotein (AFP) before and after the procedure. The frequency of fetomaternal transfusion depends on the test employed to diagnose this complication and also on placental location, since fe-tomaternal transfusion is more common with anterior placentae. The AFP method is more sensitive than the Klei-hauer-Betke test. For example, Lele et al⁵² reported that the frequency of fetomaternal transfusion was 1.8% and 7%, as detected with the Kleihauer-Betke test and the AFP method, respectively. It has been suggested, however, that in some cases the rise of AFP concentration in maternal serum after the procedure is due to contamination with amniotic fluid and not fetomaternal transfusion. This is important because fetomaternal transfusion can lead to isoimmunization (see section Isoimmunization) and an increased frequency of pregnancy loss. Mennuti et al⁵³ reported that spontaneous abortion occurred more frequently in patients with an elevation of maternal serum AFP after amniocentesis than in those with no elevation (14.2% vs 0.98%).

Recently, detection of circulating fetal nucleic acids in maternal blood by use of real-time quantitative polymerase chain reaction (qPCR) has been proposed as a promising tool for noninvasive prenatal diagnosis, as well as for quantifying fetal-maternal hemorrhage after amniocentesis. It has been found that amnio-centesis is associated with a significant increase in the fetal DNA concentrations in maternal blood, which may represent transfer of either fetal cells or fetal DNA to the maternal circulation.

Discolored Amniotic Fluid

Brown- and green-stained amniotic fluids are occasionally obtained in mid-trimester amniocenteses.^{54,55,56, and 57} Whereas brown amniotic fluid is considered an indicator of an intra-amniotic hemorrhage, green fluid has been attributed to meconium staining or an old hemorrhage. Hankins et al⁵⁷ detected total and fetal hemoglobin in brown, green, and clear second trimester amniotic fluid samples, with higher concentrations in brown amniotic fluid, and reported that fetal hemoglobin was determined to be 20% to 100% of the total hemoglobin. This evidence suggests that both green and brown amniotic fluids probably represent the occurrence of previous intra-amniotic bleeding.⁵⁷ Indeed, both green- and brown-stained amniotic fluids are positive for free hemoglobin. Furthermore, they have a similar spectrophotometric pattern that is consistent with the presence of oxyhemoglobin and free hemoglobin.

It has been suggested that the color of the fluid correlates with the amount of hemoglobin present. In vitro ex-periments, in which amniotic fluid is contaminated with blood, have indicated a sequential color change. Green-colored and brown-colored amniotic fluids were seen after 3 and 7 days of incubation, respectively. This interpretation is consistent with the clinical observation that women with green or brown amniotic fluid have a positive history of vaginal bleeding, but little is known about the etiology of the bleeding episode. Cassell et al⁵⁸ reported the recovery of Mycoplasma hominis and Ureaplasma urealyticum in 4 of 33 samples with discolored, second trimester amniotic fluid. Similarly, Gray et al⁵⁹ found the amniotic fluid to be discolored in 4 of 8 amniotic fluid samples positive for U urealyticum. It is possible that an intrauterine infection may lead to bleeding, or that bleeding and clot formations provide an adequate nidus for microbial growth. Gomez et al⁶⁰ proposed that a subclinical intrauterine infection can cause deciduitis and decidual bleeding, which can be clinically manifested as vaginal bleeding. Indeed, microbial invasion of the amniotic cavity has been implicated as a common etiology of vaginal bleeding as it was found in 14% of pregnant women with "idiopathic" vaginal bleeding. Moreover, both microbial invasion of the amniotic cavity and vaginal bleeding were associated with subsequent preterm prelabor rupture of membranes (PROM) and early preterm delivery. It has been proposed that subclinical intrauterine infection can cause deciduitis and decidual bleeding, leading to vaginal bleeding and subsequent adverse pregnancy outcomes.⁶⁰ Recently, Vaisbuch et al^61,62 measured amniotic fluid concentrations of total and fetal hemoglobin in normal pregnant women, as well as in patients with spontaneous preterm labor with intact membranes and in those with preterm PROM. The authors found that, in normal pregnancy, the median amniotic fluid total and fetal hemoglobin concentrations at term were significantly higher than that in the mid-trimester, and that women at term in labor had higher median hemoglobin concentrations than those not in labor. Among patients with preterm labor and those with pre-term PROM, no differences were found in the amniotic fluid concentrations of fetal hemoglobin in the presence or absence of intra-amniotic infection/inflammation (IAI). In contrast, the median total hemoglobin concentrations in amniotic fluid were significantly higher in patients with IAI than those without IAI. These observations suggest that the main source of fetal hemoglobin in cases with IAI is of maternal origin, and that the detected fetal hemoglobin may represent the normal percentage of hemoglobin F in the maternal blood. The fact that total hemoglobin was present in the amniotic fluid of normal and complicated pregnancies suggests that detection of hemoglobin in amniotic fluid should not be necessarily considered as an abnormal finding. Of note, no difference was found in the median total hemoglobin concentrations between women with and without transplacental amniocentesis (P = .3).

The existence of free hemoglobin in amniotic fluid and its association with brownish coloration of the amniotic fluid or clinical outcomes has been previously reported,^{54,57,63,64, and 65} although there is conflicting evidence regarding the presence of hemoglobin in clear amniotic fluid.^{57,63,64, and 65} Francoual et al⁶⁴ measured (by colorimetry) total hemoglobin concentration in amniotic fluid obtained by amniocentesis or from the vaginal pool in 78 patients in the late third trimester. A higher mean total hemoglobin concentration was found in "meconium-stained" amniotic fluid than in "clear" amniotic fluid, although the authors reported that hemoglobin was present in most samples of "clear" amniotic fluid.⁶⁴ In contrast, Legge⁶³ examined 208 second trimester amniotic fluid samples for pigmentation (visually and spectrophotometrically), and reported that none of the clear amniotic fluid samples contained abnormal pigmenta-tion, while 14 of the 15 dark brown–colored samples had chemically detectable hemoglobin (14 had hemoglobin A, and 8 also had hemoglobin F).

The possibility that some green-stained amniotic fluid could be due to meconium passage cannot be ruled out entirely. The human fetus produces and can pass meconium before the 20th week of gestation. Although the evidence is not consistent, most studies examining the prognostic significance of dark-stained amniotic fluid have suggested an increased risk of pregnancy loss. The relative risk for spontaneous abortion after retrieval of discolored amniotic fluid has been reported to be 9.9 by Tabor et al.⁴⁷ King et al⁶⁶ suggested that the prognosis is worse if discolored fluid is associated with an elevated maternal serum AFP determined before the amniocentesis. These findings are at variance with those of Hankins et al,⁵⁷ who did not find an increased frequency of poor pregnancy outcome after examining data from 83 patients with dark (n = 6) or green fluid (n = 77) of 1227 women undergoing mid-trimester amniocenteses.

Maternal Risks

Maternal complications are extremely rare. They include perforation of intra-abdominal viscera with subsequent intra-abdominal infection, bleeding, and blood group sensitization. One case of amniotic fluid embolism has been reported after a third trimester amniocentesis. The patient presented at 32 weeks of gestation with polyhydramnios. After draining 200 cc, the patient developed respiratory distress and disseminated intravascular coagulation. The mother was treated with exchange transfusions and survived, but the fetus died in utero. Severe hemorrhage due to laceration of the inferior epigastric vessels has also been reported after a third trimester amniocentesis. Thorp et al⁶⁷ reported a maternal death after an uncomplicated second trimester amniocentesis. The patient died from Escherichia coli sepsis and disseminated intravascular coagulation 40 hours after the amniocentesis. The autopsy showed normal-appearing needle entries into the skin and uterus without evidence of bowel adhesion or a needle track through the bowel. The amniotic fluid specimen demonstrated negative leukocyte esterase activ-ity, negative Gram stain for bacteria and white blood cells, and normal glucose and interleukin-6 (IL-6) concentra-tions. Similary, Kim et al⁶⁸ reported a case of a 40-year-old female patient admitted with the diagnosis of fetal death and complaining of fever, chills, progressive dyspnea, hematuria, and abdominal pain 2 days after an amniocentesis was performed at 17 weeks of gestation. The patient developed septic shock and disseminated intravascular coagulation, and was further complicated by an acute myocardial infarction. The maternal blood culture was positive for E coli. The patient survived and recovered approximately 3 weeks after the amniocentesis. Erez et al⁶⁹ reported 2 cases of septic abortion diagnosed shortly after uncomplicated amniocenteses performed under sterile technique and continuous ultrasound guidance at 18 and 24 weeks of gestation. The first patient had a history of ulcerative colitis in remission and a total proctocolectomy with an ileal pouch performed 10 years earlier, while the medical history of the second patient was unremarkable. Of interest, both patients developed fever just minutes or a few hours after the amniocentesis. The amniotic fluid cultures were positive for E. coli in the first case, and E. coli as well as Staphylococcus epidermidis in the latter. The authors propose that contamination of the uterine cavity by gastrointestinal flora due to unintended intestinal puncture during needle insertion is a plausible explanation for the first case because of the history of intestinal surgery, and that contamination of the uterine cavity after a sterile amniocentesis technique may also occur by direct inoculation of skin flora, which may be the explanation for the amniotic fluid culture positive for S. epidermidis in the second case.⁶⁹

Fetal Risks

Fetal complications can be grouped into fetal loss and needle injuries. Fetal loss can be idiopathic or due to direct fetal injury resulting in exsanguination or infection. The term idiopathic refers to unexplained fetal death that occurs during the procedure, and in which postmortem examination yields no demonstrable reason for the demise. In these cases, fetal heart activity is detected before but not after the amniocentesis. A neurogenic mechanism has been postulated, but there is no evidence to support this mechanism.

Fetal Loss

Ager and Oliver⁷⁰ analyzed the results of 28 different reports of mid-trimester amniocenteses published between 1977 and 1985. The total fetal loss rate [(total spontaneous abortions + stillbirths + neonatal deaths)/effective total number of pregnancies] differed considerably across the studies (range: 2.4% to 5.2%) [Note: The effective total number of pregnancies: total number of pregnancies − (deaths prior to amnio-centesis + total elective abortions + pregnancies lost to follow-up)]. However, the authors concluded that these estimations of risk are questionable because of differences in amniocentesis procedures, patient popula-tions, methods of risk estimation, and problems in the design and analysis of the different studies. Most studies do not have a control group, or the control group was the result of matching. Nonrandomized studies are susceptible to bias. Furthermore, most studies do not have the adequate sample size to detect a significant difference in the risk associated with the procedure.

The traditional rate of pregnancy loss related to amniocentesis recommended by the Centers for Disease Control and Prevention is 0.5%,⁷¹ which is different from the one that Tabor et al⁴⁷ reported in a randomized controlled trial of genetic amniocentesis in 4606 low-risk women. This study provides the best risk estimation available in the literature. Patients were invited to participate in a study in which they were randomized to have an amniocentesis (study group) or an ultrasound (control group) in the mid-trimester, and those at increased risk for chromosomal anomalies, neural tube defects, metabolic disorders, or spontaneous abortions were excluded. Amniocenteses were performed under sonographic guidance with a 20-gauge needle by a group of 5 physicians.⁴⁷ The rate of spontaneous abortion after 16 weeks of gestation was higher in the group that underwent amniocentesis than in the group that had an ultrasound (1.7% vs 0.7%, P < .01). The excess spontaneous abortion rate of 1% (95% confidence interval [CI], 0.3 to 1.5) corresponds to a relative risk of 2.3 (95% CI, 1.3 to 4.0). There was a different distribution of the time elapsed between procedure and spontaneous abortion in the study and control groups. The median time interval was 21.5 days (range 5 to 67) in the study group versus 46.5 days (range 8 to 70) in the control group. An elevated maternal serum AFP (greater than 2 multiples of the median for gestational age), perforation of the placenta, and dis-colored amniotic fluid were identified as risk factors for spontaneous abortion. Their relative risks for fetal loss were 8.3 (95% CI, 2.4 to 19.8), 2.6 (1.3 to 5.4), and 9.9 (4.3 to 22.6), respectively. The researchers pointed out that the 1% increased risk of spontaneous abortion after mid-trimester amniocentesis may be an underestimation of the real risk. Termination of pregnancies with fetuses affected with chromosomal abnormalities (identified in the study group and not in the control group) may have artificially reduced the rate of spontaneous abortion of the study group. Contrary to what has been reported in previous studies, no correlation was found between the rate of spontaneous abortion and number of needle insertions, placental site, or experience of the operator. Whereas amniotic fluid leakage occurred more commonly in the study group than in the control group (1.7% vs 0.4%, P < .001), vaginal bleeding occurred with similar frequency (2.4% vs 2.6%). Other obstetric complications such as preterm delivery, spontaneous rupture of membranes, and abruptio placentae had similar prevalence in both groups.

A randomized trial of amniocentesis compared with no amniocentesis to try to define contemporary procedure-related loss rates is not likely to be performed due to feasibility and ethical considerations. Eddleman et al⁷² attempted to quantify the contemporary procedure-related loss after mid-trimester amniocentesis using a database generated from patients who were recruited to the First and Second Trimester Evaluation of Risk (FASTER) for Aneuploidy trial, which included 35,003 patients who either did (study group, n = 3096) or did not undergo mid-trimester amniocentesis (control group, n = 31,907). The rate of fetal loss before 24 weeks of gestation was compared between the 2 groups. The observed difference in crude pregnancy loss rates less than 24 weeks between the groups was 0.06% (1.0% minus 0.94%, respectively). This study reveals an amniocentesis-related loss risk of approximately 1 in 1600, which is substantially lower than the traditionally quoted risk of 1 in 200.⁷¹ Despite large number of patients included, the nonrandomized design, limited information concerning the needle size used for the procedure, and the experience of clinicians performing the procedures are considered as potential biases of this study. Odibo et al,⁷³ in a retrospective cohort study including all pregnant women presenting for prenatal care between 1990 and 2006 at the Department of Obstetrics and Gynecology Washington University, compared the fetal loss rate of women who underwent amniocentesis between 15 and 22 weeks of gestation (n = 11,746) with those who did not have any invasive procedure and had a live fetus documented on ultrasound examination between 15 and 22 weeks (n = 39,811). The fetal loss rate less than 24 weeks attributable to amniocentesis was 0.13% (1 of 769). It is possible that increased quality image of the current ultrasound machines and experience of clinicians performing amniocentesis may account, in part, for the reduction in the complication rate associated to amniocentesis in the last decade.

The Practice Bulletin of the American College of Obstetricians and Gynecologists published in 2007⁷⁴ states that early amniocentesis (procedure conducted at less than 15 weeks of gestation) should not be performed because of the higher risk of pregnancy loss and complications compared with traditional amniocentesis (procedure conducted at 15 weeks of gestation or later). It also states that amniocentesis at 15 weeks of gestation or later is a safe procedure. According to the Bulletin, the procedure-related loss rate after mid-trimester amniocentesis is less than 1 in 300 to 500.⁷⁴

Recently, an excellent review conducted by Seeds⁷⁵ included 29 studies published between 1976 and 2002 that reported at least 1000 mid-trimester amniocenteses each, as well as information about the spontaneous pregnancy loss risk after the procedure and the use of ultrasound guidance during amniocentesis. The author focused on the rate of spontaneous pregnancy loss between the procedure and 28 weeks of gesta-tion. Patients were excluded if they were lost to follow-up, or had therapeutic abortions, third trimester losses, or perinatal deaths. A total of 68,119 mid-trimester genetic amniocenteses were included and divided into 2 groups according to the performance of amniocentesis with or without ultrasonographic guidance. The first group included 33,975 amniocenteses with ultrasound only before the procedure. The overall risk of pregnancy loss before 28 weeks of gestation was 2.1%. In contrast, 34,144 amniocenteses were performed under the sonographic monitoring technique with a spontaneous pregnancy loss of 1.4%. Further, when the analysis was restricted only to studies that included control patients, the rate of pregnancy loss between amniocentesis with concurrent ultrasound monitoring and control subjects was significantly different (1.7% vs 1.1%, respectively; P < .0001), with a procedure-related rate of fetal loss of 0.6%. Finally, the author reported 9 studies that addressed the impact of transplacental amniocentesis on the risk of pregnancy loss. The aggregate rate of spontaneous pregnancy loss was 1.4% (73 of 5203), similar to the overall pregnancy loss rate after ultrasound-monitored amniocentesis and lower than that reported from controlled studies, suggesting that transplacental amniocentesis in the mid-trimester does not increase the rate of pregnancy loss.⁷⁵

Infection

Amniocentesis may lead to intra-amniotic infection by introducing microorganisms into the amniotic cavity (ie, contaminated instruments, passage of the needle through contaminated skin, or intra-abdominal viscera). Alternatively, if amniocentesis results in the rupture of membranes, ascending infection may occur. Although blood cultures obtained around the time of the procedure have been negative in a small study,⁷⁶ and antibiotic pro-phylaxis is not a routine practice, the mid-trimester period seems to be particularly vulnerable to microbial invasion as the antibacterial activity of amniotic fluid is at its nadir.⁷⁰ Some case reports have suggested a temporal relationship between the procedure and clinical chorioamnionitis; however, the prevalence of intra-amniotic infection after mid-trimester amniocentesis is unknown.

The traditional view is to consider intra-amniotic infection as an acute complication of pregnancy. Yet, there is evidence implicating infection as an etiologic factor for pregnancy loss after mid-trimester amniocentesis. Evidence from studies of the microbiological state of the amniotic cavity using traditional culture techniques or polymerase chain reaction (PCR) at the time of genetic amniocentesis in asymptomatic patients suggest that, in some cases, intra-amniotic infection may precede the procedure rather than follow it. Cassell et al⁵⁸ were the first to report the presence of genital mycoplasmas in 6.6% (4 of 61) of amniotic fluid samples collected by amnio-centesis between 16 to 21 weeks. Two patients had positive cultures for M hominis and 2 for U urealyticum. Patients with M hominis delivered at 34 and 40 weeks without neonatal complications, whereas those with U urealyticum had premature delivery, neonatal sepsis, and neonatal death at 24 and 29 weeks. In a series of 2641 second trimester genetic amniocenteses reported by Gray et al,⁵⁹ all of which were cultured for mycoplasmas, 0.37% (9 of 2461) had a positive culture for U urealyticum. One patient was excluded from the analysis because of a subsequent therapeutic abortion. The perinatal outcome of the other 8 patients was compared with 86 patients with complete follow-up having genetic amniocenteses during the same study period and negative cultures for mycoplasmas. Among the 8 patients who cultured positive for U urealyticum, 75% (6 of 8) had a spontaneous abortion within 4 weeks of the procedure as compared with only 1.2% (1 of 86) of the patients with negative cultures. The other 2 patients in the positive culture group delivered prematurely at 24 and 30 weeks, and only 1 of the infants survived. All 8 placentas had evidence of chorioamnionitis on histologic examination. Fifty percent (4 of 8) of the samples obtained from Ureaplasma-positive patients were discolored, while only 2.3% (2 of 86) of the fluid samples obtained from Ureaplasma-negative patients. Discolored amniotic fluid was significantly associated with the presence of U urealyticum in amniotic fluid (P < .001) and an adverse perinatal outcome. Horowitz et al⁷⁷ detected U urealyticum in 2.8% (6 of 214) of amniotic fluid samples obtained between 16 and 20 weeks of gestation. The rate of subsequent adverse pregnancy outcome (fetal loss, preterm delivery, and low birth weight) was significantly higher in patients with a positive amniotic fluid culture than in those with a negative culture (50% vs 12%, P = .035). Simi-larly, Berg et al⁷⁸ found 1.6% (44 of 2718) positive cultures for U urealyti-cum/M hominis. Among these 44, 35 patients were treated with oral erythromycin. Mid-trimester pregnancy loss was significantly higher in the untreated than in the treated group (44% vs 11%; P = .04).

Fetal Injuries Associated with Amniocentesis

Table 25-2 presents fetal injuries that have been attributed to amniocentesis and whether or not ultrasound was employed during the procedure. The spectrum of lesions ranges from mild skin dimples to fetal death due to exsan-guination.^{79,78,79,80,81,82,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104, and 105} Although fetal injuries are generally associated with bloody taps, they have also been reported after clear taps. If a fragment of tissue is retrieved, histologic examination is recommended to establish its origin. Fetal injuries have occurred even with the use of a sonographically monitored technique. The most frequent lesion associated with amniocentesis is skin puncture. Although a cause-effect relationship is difficult to estab-lish, needle injuries should be suspected if the shape of the lesion resembles a needle track or a depressed punctiform scar.

Several ocular injuries have been attributed to amniocentesis. Typically, these are unilateral lesions detected shortly after birth. In one case,⁹⁷ the newborn had a small and cloudy eye, a coloboma of the upper lid, and a hazy and edematous cornea. The combination of lesions in the eyelid and cornea suggests that the injury occurred before separation of the eyelids. The mother had an amniocentesis at 19 weeks and the first 2 cc of fluid were bloody, but the subsequent 30 cc were clear. In 2 cases,^82,86 a red and photophobic eye in the newborn period was subsequently associated with the development of an enlarging cystic mass in the anterior chamber of the eye. The cystic lesions evolved over a period of several months and were lined by a stratified squamous epithelium. Similar findings have been reported for 2 children with unilateral and progressively large epithelial iris cysts occupying nearly half of the pupil. The lesions were diagnosed at 8 months and 5 years of age, respectively, and both mothers had an amniocentesis performed at 18 and 43 weeks. The children had no history of postnatal ocular trauma.¹⁰⁶ Five other children were reported with lesions attributed to ocular perforation during amniocentesis: the first child had an amniocentesis at 16 weeks and was found at birth to have a cystic lesion communicating with the right lateral ventricle, left homonymous hemianopia, and possible damage to the right optic tract; the second child had an amniocentesis at 30 weeks and presented in the neonatal period with a distorted pupil toward the 3 o'clock position and a small tag or iris drawn up toward the full-thickness corneal scar; the third child had an amniocentesis performed at 16 weeks and presented at birth with left exotrophia, limitation of abduc-tion, and microphthalmia; the fourth child had an amniocentesis performed at 15 weeks and was noted at 31/2 years to have a small adherent leukoma near the limbus; the last child in this series had an amniocentesis performed during the second trimester and was found at 5 years of age to have a small chalazion on its right upper lid with a small full-thickness scar near the limbus.¹⁰⁷

A porencephalic cyst in a newborn with 2 subcutaneous nodules in the right and left occipital region (suggesting a needle track) has been reported.⁹⁸ An amniocentesis was performed at 18 weeks. In addition, in utero injection of contrast in the ventricular system during the course of amniograms has been reported. Squier et al¹⁰⁸ reported 5 cases of brain injury after mid-trimester amniocenteses (16 to 18 weeks). In the first case, an amniocentesis was performed under ultrasonographic guidance. The tip of the needle was not seen after its insertion and the fetus moved to the target area. The needle was found to be inserted into the fetal brain. Tissue fragments found in the amniotic fluid corresponded to germinal matrix and choroid plexus. Follow-up ultrasounds were considered normal but the mother decided to terminate the pregnancy. The brain was examined and a subarachnoid hemorrhage at the base of the brain and over the left sylvian fissure was observed, in addition to hemorrhage and necrosis of the deep white matter, lentiform nucleus, and internal capsule. The second and third cases corresponded to amniocenteses performed without ultrasonographic guidance. Both infants developed seizures, cerebral palsy, and developmental delay. In the second case, there was atrophy of the right cerebral hemisphere and a hemispherectomy was performed at 8 years of age. The fourth case was a twin pregnancy that had numerous needle insertions at 16 and 17 weeks of gestation under sonographic guidance. One twin had a skin dimple and atrophy of both left cerebral and cerebellar hemispheres, developed cerebral palsy, and died at 7 years of age due to respiratory complications secondary to cerebral palsy. Finally, the fifth case was found to be hypotonic at the time of delivery and had a skin scar in the occipital region. The patient had an amniocentesis performed without ultrasonographic guidance at 16 weeks of gestation. The infant developed cerebellar hypoplasia and cerebral atrophy, and died because of a broncopneumonia at 2 years of age.¹⁰⁸

Thoracic lesions associated with amniocentesis include hemothorax, pneumothorax, and fetal cardiac tampo-nade. In the abdomen, injuries have ranged from laceration of the liver, kidney, and spleen to ileocutaneous fistula with ileal atresia. In one case, a fragment of tissue retrieved during amniocentesis grew small bowel mucosa in culture, confirming intraoperative bowel injury.¹⁰⁰

Limb lesions have included disruption of the patellar tendon, gangrene of one arm (perforation of the subclavian artery), and an arteriovenous fistula between the popliteal artery and vein. Amniocentesis has been implicated by some investigators in the etiology of amniotic band syndrome; however, there is no agreement regarding a cause-effect relationship between this syndrome and amniocentesis. Two cases of hematomas of the umbilical cord have been reported.^82,104 Fetal exsanguination due to vascular puncture has also been reported.

Other Complications

The frequency of orthopedic congenital anomalies was not different in the amniocentesis and control groups in the study conducted by Tabor et al.⁴⁷ These results are comparable with the findings of a case control study in which amniocentesis had not been performed more often in mothers of newborns with orthopedic abnormalities than in a control group.¹⁰⁹ In contrast, the British⁵⁰ and Ameri-can⁴⁶ collaborative studies reported an increased incidence of orthopedic abnormalities (talipes equinovarus, congenital dislocation of the hip, and metatarsus abductus) in newborns of mothers who underwent amniocentesis.

Respiratory distress syndrome is considered a potential complication of amniocentesis. Indeed, the prevalence of respiratory distress syndrome was higher in neonates born to mothers in the study group than in those born to mothers in the control group (1.1% vs 0.5%, P < .05).⁴⁷ A similar finding was reported for neonatal pneumonia (0.7% vs 9.3%, P < .05). These observations are of considerable interest because they are consistent with those of the British Collaborative Study⁵⁰ in which there was an increased incidence of respiratory distress (defined as respiratory difficulties requiring oxygen and lasting more than 24 hours) in neonates born to mothers who had undergone amniocentesis in comparison with those in the control group (1.27% [30 of 2370] vs 0.38% [9 of 2402]).⁴⁷ In addition, the mean crying vital capacity, a measure of lung volume, was found to be lower in 10 neonates of mothers undergoing mid-trimester amniocentesis in comparison with those in a control group. Studies in monkeys (Macaca fascicularis), specifically designed to study the effect of amniocentesis on lung development, have indicated that a reduction in the number of alveoli and in lung volume can occur after amniocentesis at a period equivalent to 14 to 17 weeks of gestation in humans. Thompson et al¹¹⁰ evaluated the prevalence of respiratory distress and lung growth by measuring the functional residual capacity (FRC) in 74 newborns of mothers who had an amniocentesis at 10 to 13 weeks, and in 86 newborns of mothers who had a CVS during the same gestational age interval. Six infants in the CVS group, but none in the amniocentesis group, required admission to the intensive care unit because of respiratory distress (P < .05). The FRC was not different between the 2 groups. The overall incidence of FRC values below the 2.5th percentile was higher than expected (9%), indicating that both amniocentesis and CVS performed in the first trimester of pregnancy may impair antenatal lung growth. Although other clinical studies have not demonstrated an increased incidence of pulmonary complications after am-niocentesis, their design and sample size are not as adequate as those reported by Tabor et al.⁴⁷

Leakage of Amniotic Fluid

Transient vaginal leakage of small volumes of amniotic fluid after genetic amniocentesis occurs in approximately 1% to 2% of all cases and resolves spontaneously within 48 hours. Devlieger et al¹¹¹ reported a prevalence of 4% of vaginal fluid loss in 100 amniocenteses performed between 14 and 20 weeks of gestation. Tabor et al⁴⁷ reported amniotic fluid leakage in 1.7% of patients (39 of 2239). In a study that included 3469 early amniocenteses (11.7 to 14.9 weeks), 28% were performed at 12 weeks, 51% at 13 weeks, and 21% at 14 weeks. The rate of leakage of fluid decreased with advancing gestational age (2.5%, 1.7%, and 1%, respectively).¹¹² Chronic leakage of amniotic fluid is rare. Of the 8 cases reviewed by Crane and Rohland,¹¹³ preterm delivery before 32 weeks occurred in 3 cases, while clubfoot deformity in 2 cases. One death occurred in an infant delivered at 31 weeks of gestation with clubfoot deformity and Potter fa-cies. Although there are not enough data to support definitive conclusions, expectant management was associated with successful pregnancy outcome in 6 of 7 patients who experienced amniotic fluid leakage within 24 hours of a genetic amniocentesis in a series of 603 patients reported by Gold et al.¹¹⁴ All patients were placed on bed rest, had digital cervical examination prohibited, frequent white blood cell counts with differential analysis, and close maternal surveillance for clinical evidence of chorioamnionitis. Six patients delivered healthy neonates at term, and 1 patient had an intrauterine fetal demise at 25 weeks. In this case, an unsuccessful amniocentesis was attempted 6 weeks before delivery.

Long-Term Outcome

Mid-trimester amniocentesis has been used as a diagnostic procedure for more than 20 years. A Canadian trial¹¹⁵ compared the long-term outcome (7 to 18 years) of 1297 children whose mothers had mid-trimester amniocentesis for advanced maternal age with a group of 3704 controls matched for maternal age, sex, date, and place of birth. With the exception of a higher risk of ABO isoimmunization (9.5% [6 of 1297] vs 0.03% [1 of 3704]; relative risk 3.3; 95% CI, 2.4 to 4.5), the offspring of women who had mid-trimester amniocenteses did not have more disabilities than those who did not (including cerebral palsy, delayed speech, intellectual impair-ment, hearing deficits, epilepsy, asthma, and limb anomalies).

Isoimmunization

Fetal red blood cells contain the D antigen on their surface and are capable of immunizing the Rh-negative mother after a fetomaternal transfusion in the mid-trimester. This event can occur spontaneously during pregnancy or after an amniocentesis. The World Health Organization (WHO) and the ACOG have recommended the administration of the anti-D immunoglobulin G (IgG) to women after mid-trimester amniocentesis. There is no agreement on the dose; the WHO recommends 50 μg, whereas the ACOG recommends 300 μg.¹¹⁶ The basis for this recommendation is that mid-trimester amniocentesis has been associated with an increased incidence of transplacental hemorrhage, a risk factor for isoimmunization, although the precise risk of isoimmunization after mid-trimester amniocentesis has not been well defined. The incidence of Rh isoimmunization in the randomized controlled clinical trial reported by Tabor et al⁴⁷ was 0.3% (7 of 370) in the study group and 0.1% (3 of 347) in the control group (anti-D IgG was not administered to Rh-negative patients undergoing amniocentesis). Although this difference is not significant, the number of patients required to detect a difference of 1% between the amniocentesis and the controls group would be 2896 in each group.⁴⁷ The analysis of the previous reports suggests that mid-trimester amniocentesis is associated with an increased risk of isoimmunization, and the magnitude of the increased risk seems to be approximately 1%.⁷⁰

Murray et al¹¹⁷ provided a comprehensive analysis of the advantages and disadvantages of anti-D IgG administration. The objections that have been raised against the routine use of anti-D IgG are unproven efficacy (isolated case reports have indicated that sensitization can occur after anti-D IgG administration) and unproven long-term safety. Anti-D IgG crosses the placenta and coats Rh-positive fetal red cells. It is unclear if this could have adverse effects. The theoretic risk of augmentation has been suggested. This phenome-non consists of an enhancement of the immune response in the context of small amounts of antibody. Further-more, the long-term effects of exposing the immunologically "naïve" fetal immune system to human immunoglobulins are unknown. Although there is no irrefutable evidence to support the routine administration of anti-D IgG after mid-trimester amniocentesis, it has become the standard practice in the United States.

Traditionally, genetic amniocentesis has been carried out after the 16th week of gestation. To obtain results ear-lier, several centers have introduced the performance of "early amniocentesis" at a gestational age lower than 15 weeks. Early amniocentesis is an attractive approach because it provides early cytogenetic diagnosis and amniotic fluid for AFP determinations for the assessment of neural tube defects.^20,24,32 Wald and Cukle¹¹⁸ demonstrated that amniotic fluid AFP obtained between 13 and 15 weeks had a 100% sensitivity for the diagnosis of anencephaly and a 96% sensitivity for the diagnosis of spina bifida when using cut-off values greater than 2.0 and 2.5 multiples of the normal median, respectively.

Although early amniocentesis has been performed as early as the seventh week of gesta-tion,³² it is usually performed between 11 and 15 weeks of gestation. The technique for early amniocentesis is basically the same as for mid-trimester amniocentesis, with minor modifications: (1) most investigators have recommended the use of a smaller bore (22-gauge) spinal needle;^{15,16,17,18,19,20,21,22,23,24,25,26,27, and 28,31,32,35} (2) the fluid should be aspirated slowly to prevent collapse of the amniotic sac;²⁶ and (3) because the amniotic membrane may not be completely fused with the chorion at this state in pregnancy, membrane tenting is a more frequent complication than with mid-trimester amniocentesis.

Transvaginal aspiration of amniotic fluid has been proposed as an alternative to the transabdominal approach for early amniocentesis. Potential advantages include the high resolution of the transvaginal probe and easy access to the amniotic sac. Jorgensen et al³² attempted the procedure in 36 women between 7 and 12 weeks of gestation. Although amniotic fluid was obtained in all cases and the patients tolerated the procedure well, culture was unsuccessful in 6 cases (16.7%) because of bacterial or fungal overgrowth. In contrast, culture success was reported in all 96 control samples obtained by the transabdominal route. Shalev et al¹¹⁹ compared the clinical and laboratory results of first trimester transvaginal amniocentesis with those of transcervical CVS and mid-trimester amniocentesis. Transvaginal amniocentesis was performed in 355 women between 10 and 12 weeks of gestation using a 20-cm-long 22-gauge needle. The volume of amniotic fluid retrieved was 1 mL per week of gestation. Three hundred fifty-six consecutive transcervical CVS and 356 consecutive mid-trimester transabdominal amniocenteses were matched for maternal age and indication for the procedure and selected as controls. Amniotic fluid was successfully retrieved in 99.7% (355 of 356) of the first trimester amniocenteses and 100% (356 of 356) of the mid-trimester amniocenteses. In comparison, CVS was successful in only 97.8% (346 of 356) of the cases (P < .05 vs first trimester and mid-trimester amniocen-teses, respectively). No significant differences in culture success rates were observed between patients undergoing early transvaginal amniocentesis (97.9% [344 of 355]), mid-trimester transabdominal amniocentesis (97.2% [348 of 356]), and CVS (96.5% [335 of 346]). The spontaneous pregnancy loss, however, was significantly higher in patients undergoing early transvaginal amniocentesis than in patients who had mid-trimester transabdominal amniocentesis (3.2% [11 of 345] vs 0.9% [3 of 350], P < .05). There were no differences in total pregnancy loss between either group or patients who had CVS (2.9% [10 of 344]).¹¹⁹

The major concern with early amniocentesis is the increased risk of spontaneous fetal loss in comparison with either CVS or conventional amniocentesis. The Canadian Early and Mid-Trimester Amniocentesis Trial (CEMAT) Group¹²⁰ conducted a randomized trial in 4374 patients to assess safety and efficacy of early and mid-trimester amniocentesis. The mean gestational age and amniotic fluid volume retrieved at early amniocentesis were 12.1 weeks and 11 cc, respectively, while in patients that underwent mid-trimester amniocentesis they were 15.4 weeks and 20 cc, respectively. After excluding intrauterine and neonatal death, the group reported a post-procedure spontaneous loss rate of 2.6% and 0.8% for early and mid-trimester amniocentesis, respectively. When only patients who received their procedure within the allocated window were considered (<13 weeks and >15 weeks), the post-procedure spontaneous loss rate for early and mid-trimester amniocentesis was 2.7% and 0.5%, respectively.⁴⁵ The procedure-related variables that were found to be associated with pregnancy losses after early amniocentesis were (1) difficult procedures (5.7% vs no difficult procedure: 2.4%, P = .001); (2) amniotic fluid leakage (11.7% vs no amniotic fluid leakage: 2.4%, P = .001); and (3) vaginal bleeding (10.5% vs no bleeding: 2.1%, P = .001). Interestingly, although a trend towards lower rates of pregnancy loss among the most experienced operators and centers was observed, the differences were not statistically significant. This suggests that operator and center experience were not associated with a higher rate of pregnancy loss.

Early amniocentesis was associated with a higher rate of multiple needle insertions (5.4% vs 2.1%, P < .0001) and amniotic fluid leakage before 22 weeks of gestation (3.5% vs 1.7%, P = .0007) than that of mid-trimester amniocentesis. The incidence of talipes equinovarus was higher in the early amnio-centesis group than in the mid-trimester group (1.6% vs 0.1%, P < .001).⁴⁵ A possible explanation for talipes is the removal of a relatively large amount of amniotic fluid (11 mL, about 20% of the total amount of amniotic fluid) during early amniocentesis in the CEMAT study. Cytogenetic results from the CEMAT study (including culture success, detection of chromosome abnormalities, frequency of mosaicism, and maternal cell con-tamination) showed no difference between early and mid-trimester amniocentesis. A repeat amniocentesis was re-quired more frequently after early amniocentesis than after mid-trimester amniocentesis.¹²¹

Table 25-3 displays the overall spontaneous pregnancy loss of early amniocentesis reported in published studies including at least 100 cases.^{120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138, and 139} The fetal loss rate is related to the gestational age at which the procedure is performed. Table 25-4 shows the data of 5 different studies in which the fetal loss rate by gestational age at amniocentesis was reported. The overall fetal loss rate was significantly higher when the am-niocentesis was performed before 12 weeks of gestation than after this gestational age (8.1% [5 of 62] vs 1.3% [38 of 2950], P < .001).^{16,20,26,30,31} The combined experience of 7 centers for which data are available (Table 25-5) shows an overall loss rate of 1.3% (44 of 3296) within 14 days of the procedure, with a significant increase when amniocentesis is performed before 12 weeks (6.9% [5 of 72] vs 1.2% [39 of 3224], P < .001). ^{16,20,23,26,28,30,31,34}

In assessing the ability to culture amniocytes, Nelson and Emery¹⁴⁰ demonstrated that the largest percentage of viable cells was present between 13 and 16 weeks of gestation. Several studies have reported a higher culture failure rate and a longer time to reach a final diagnosis after early amniocentesis than after mid-trimester amniocentesis.^{20,124,141,142, and 143} This has been attributed to a lower cell concentration in the amniotic fluid.²⁰ Moreover, confined mosaicism may be more likely in fluid retrieved after an early amniocentesis. Up to 70% of cells in amniotic fluid during the first trimester are derived from the membranes and trophoblast rather than from the fetus. With the use of a new filter technique some problems of early amniocentesis have been, in part, overcome. In 1995, Sundberg et al¹²⁹ reported the pregnancy outcome for 249 patients subjected to an early amniocentesis with a filtration technique. Sampling failure occurred in 8 cases (3.2%), and spontaneous losses before 28 weeks occurred in 5 cases (2%). A subsequent report by the same author described the use of early amniocentesis with amniotic fluid filtration in 44 pregnancies with a previous finding of mo-saicism. Amniocenteses were performed at a mean gestational age of 12.5 weeks (range of 11 to 14 weeks) after CVS had demonstrated mosaicism. In all 44 cases, culture of the filtered fluid was successful, and cultures were harvested after a mean time of 8.8 days. Mosaicism was confirmed in 7 cases, and nonmosaic karyotypes were found in 36.

The question of whether early amniocentesis is a reasonable alternative to CVS has been addressed by randomized clinical trials. In a study conducted by Nicolaides et al,¹²⁶ 488 patients were randomized to either early amniocentesis (n = 238) or CVS (n = 250) at 10 to 13 weeks. An additional group of 813 patients had either an early amniocentesis (n = 493) or CVS (n = 321) by choice. Patients who were randomized to CVS had a lower spontaneous fetal loss rate (intrauterine or neonatal deaths) than patients randomized to early amniocentesis (1.2% [95% CI, 0.3 to 3.5] vs 5.9% [95% CI, 3.3 to 9.7], P < .01). Cultures obtained by CVS demonstrated more mosaicism than cultures obtained by early amniocentesis (1.2% vs 0.1%, P < .05), although a larger number of culture failures occurred in the early amniocentesis group (2.3% vs 0.5%, P < .05). Another trial¹⁴⁴ included 115 patients that were randomized to either early amniocentesis (n = 55) or transabdominal CVS (n = 60). Seventy patients who refused randomization chose early amniocentesis, and 25 chose transabdominal CVS. Several patients did not have the procedure to which they had been assigned. Thus, 130 had early am-niocentesis and 74 received transabdominal CVS. Spontaneous fetal loss occurred in 8 patients who had early amnio-centesis (6.2%) and in none of the patients who had transabdominal CVS. Talipes equinovarus complicated 4 pregnancies that underwent early amniocentesis (3.1%) and none of the pregnancies exposed to transabdominal CVS. Sundberg et al¹⁴⁴ also compared early amniocentesis performed with a filter technique with CVS. This group also compared 555 cell cultures and karyotypes obtained by CVS with 548 obtained by early amniocentesis with a filter technique. No significant differences in culture success, pseudomosaicism, preparation problems, or maternal cell contamination were found between the groups. In addition to filtration, other technical modifications have been proposed to improve the results of early amniocentesis. For example, Eiben et al¹³³ reported an incidence of talipes equinovarus of only 0.43% when removing 3.5 mL of amniotic fluid during early amniocentesis. However, due to the high risk of complications associated to early amniocentesis (high rate of pregnancy loss, talipes equinovarus, leakage of fluid, etc), the ACOG Practice Bulletin from 2007 suggests that early amniocentesis should not be per-formed, while amniocentesis after 15 weeks is considered a safe procedure.⁷⁴

One of the most frequent indications for an invasive procedure for prenatal diagnosis is the detection of fetal aneuploidies, specifically trisomy 21 (Down syndrome), trisomy 18 (Edwards syndrome), and trisomy 13 (Patau syndrome), as well as those involving sex chromosomes such as 45,X (Turner syndrome) and 47,XXY (Klinefelter syndrome). The traditional method used for analysis of chromosomal abnormalities is karyotype, which requires intact, living cells for analysis of chromosome metaphase that is banded using Giemsa. This technique has limitations such as culture failure and external contamination, but mainly the long time required to obtain results of genetic amniocentesis, which is approximately 2 weeks. This is important, because early results from a rapid test, although they can be considered as partial results, significantly reduced the maternal anxiety during the waiting period compared with receiving only the full karyotype results (mean daily score 12.5 vs 14.8; scale score difference −2.36, 95% CI, −1.2 to −3.6; calculated using daily scores from the short version of the Spielberger State-Trait Anxiety Inventory).¹⁴⁵

Fluorescence in situ hybridization (FISH) has been proposed as an adjunctive tool for a rapid diagnosis of the most frequent chromosomal aberrations. FISH on uncultured amniocytes with the use of chromosome-specific probes has been utilized for the diagnosis of numeric abnormalities affecting chromosomes 13, 18, 21, X, and Y. Morris et al¹⁴⁶ reported the results of a prospective study using FISH on uncultured amniocytes. Of 1504 amniotic fluid samples tested, 1467 were correctly identified as having 2 copies of chromosomes 21, and 35 samples were correctly scored as having 3 copies of chromosome 21. Only 2 samples failed to produce re-sults. On average, results were reported 48 hours after sampling (range 1 to 7 days).¹⁴⁶ These findings are consistent with those of Eiben et al¹⁴⁷ who reported a correct identification of a normal karyotype in 98% (727 of 741) of cases and of an abnormal karyotype in 95.8% (68 of 71) of cases. FISH has the advantage over routine cytogenetic analysis that it requires a small amount of amniotic fluid and it provides rapid and accurate diagnosis of the most frequent chromosomal abnormalities within 24 hours of the proce-dure. This makes this method of particular value when an early amniocentesis is performed. There is, how-ever, the disadvantage that FISH can be used to detect only aberrations for which it has been designed; there-fore, most structural and some numeric chromosomal abnormalities cannot be diagnosed with this method. In ad-dition, the procedure requires considerable laboratory expertise. The American College of Medical Genetics still recommends that FISH on uncultured amniocytes be considered investigational, and clinical actions should not be based on the results of FISH alone.¹⁴⁸ However, some investigators have suggested that, in the presence of an abnormal ultrasonographic finding in the second trimester, an informative FISH result can be used to make management decisions even before the final cytogenetic result is available.

Recently, quantitative fluorescent PCR (QF-PCR) was introduced as a new molecular method for detection of chromosomal abnormalities. This method amplifies short nucleotide sequences of DNA (short tandem repeats [STRs]) of various lengths, which are chromosome specific, highly polymorphic, and stable during life. A fluoro-chrome is incorporated in the resulting amplified products to allow visualization and quantification by an automated DNA scanner. In normal heterozygous individuals, each chromosome-specific STR will produce 2 peaks of fluorescence activity with a ratio near 1:1, whereas samples from trisomic fetuses will show either 3 peaks with a ratio of 1:1:1 (trisomic triallelic) or 2 peaks with a ratio of 2:1 (trisomic diallelic). At the same time, and using appropriate STRs, selected translocations and singlegene defects such as thalassemia, sickle cell disease, and cystic fibrosis can be identified by QF-PCR. One of the advantages of this method is that it takes approximately 4 hours to perform DNA extraction and QF-PCR amplification, and the analysis of each fluorescent product is performed in 30 minutes. Moreover, automated DNA scanners can analyze single, 16, or over 90 samples simulta-neously. A limitation is that chorionic or amniotic fluid samples cannot be analyzed by QF-PCR if they are contaminated with maternal cells, which occurs in approximately 2% of cases. A degree of agreement of 96.8% has been reported between patients with normal FISH or QF-PCR, and normal karyotype. A false-negative rate of 0.9% for FISH or QF-PCR, and a false-positive rate of 1 in 214,031 for FISH have been re-ported.¹⁴⁹

Another molecular diagnostic technique for which applications in amniocentesis are developing is microarray comparative genomic hybridization (CGH) analysis. Development of a prenatal diagnostic technique to complement the aforementioned conventional karyotype and FISH for specific chromosomes would be desirable because numerous microdeletion syndromes are not detectable by these techniques, even with analysis of blood, and the resolution for prenatal karyotype is lower than that achieved for blood. Therefore, microar-ray-CGH has been proposed as a possible approach to identify chromosomal abnormalities not detected by conventional karyotype and FISH. In addition, it has been suggested that the rapid turnaround time for this technique, applied to uncultured amniocytes, could make it advantageous versus karyotype and FISH for detection of trisomy and translocations. Array-CGH has been described as a modified in situ hybridization method that can be applied to the entire genome. Areas of possible abnormality are identified by comparing DNA sequence copy number from study subjects with normal controls, and the chromosomal locations yielding the copy number variations are mapped. Over-representation of the hybridized signal (high DNA copy number) for a portion of the genome represents trisomy, tetrasomy, or gene amplification for that region, while under-representation (low DNA copy number) indicates monosomy or absence of that region.

In pediatric genetics, for example, microarray-CGH may be useful in the evaluation of children with unexplained phenotypic abnormalities, such as developmental delay, and normal conventional karyotypes. In prenatal diagnosis, an emerging application of the technique may be to serve as a follow-up analysis of an amniocentesis specimen with a normal karyotype for a pregnancy with significant ultrasound abnormalities, or eventually even as a routine test. A 2006 report of a prospective study of array-CGH at Baylor provides an illustration of the feasibility of the approach.¹⁵⁰ Couples considering invasive prenatal diagnosis were invited to undergo CGH as well. Among the 98 who agreed to participate, conventional G-banded karyotype revealed 5 abnormal pregnancies, 4 with trisomy 21, and 1 with an unbalanced translocation of chromosomes 3 and 7. Simultaneously, array-CGH was performed in a blinded fashion for the 98 subjects. The array-CGH analysis was completely concordant with the karyotype analysis for the cases of trisomy and the unbalanced translocation. In 30 cases, copy number variants were identified that were presumed to be normal, due to high frequency in normal controls. In 12 additional cases copy number variants were identified that required further analysis because they were not known to be normal or associated with a known genetic syndrome. In 9 of those 12 cases, the copy number variant was detected in a parent by array-CGH and/or FISH, and thus presumed to be benign. In 2 of the 12 cases, the copy number variants appeared to be benign on follow-up. For 1 of the 12 cases with a copy number variant not known to be normal, a gain in copy number for 2 clones of 18p were detected that were not present in parental samples. The fetus was noted to have pentalogy of Cantrell and died in the first month after birth. The copy number variant was not detected on postnatal tissue and was postulated to represent either low-level mosaicism or artifact from tissue culture.¹⁵⁰ This series illustrates both the potential utility and some of the limitations to the use of array-CGH for prenatal diagnosis at this time, with data still emerging. In a more recent re-port,¹⁵¹ a series of 300 women undergoing amniocentesis or CVS for either risk factors, such as advanced maternal age, or abnormal ultrasound findings also underwent array-CGH. Copy number variants were detected in 58 samples, of which 40 were interpreted as likely benign, 15 were associated with defined pathology, and 3 were of uncertain significance. The copy number variants described as having "defined clinical significance" included pregnancies with abnormalities such as trisomy, deletions, duplications, and ring and marker chromo-somes. In 7 of the cases, array-CGH was deemed to have contributed important new information. In 2 of the cases, abnormalities were identified that "would not have been detected without CGH." The first was a 9q34 deletion, and the second was a deletion in a region associated with thrombocytopenia absent radius syndrome (1q21) in a fetus subsequently diagnosed with absent radii.¹⁵¹ Several other reports have further described the use of array-CGH in prenatal diagnosis.

At present, the challenge for the use of array-CGH in prenatal diagnosis remains determining which copy number variants are pathologic and the effect on prognosis for an individual fetus. Although a copy number variant may be presumed to be pathologic for a patient ascertained by virtue of known diagnosis, such as developmental delay, and an otherwise normal diagnostic evaluation, the prospective significance of a copy number variant in a fetus may be uncertain. Comprehensive description of the outcomes associated with various copy number variants will be required in order to distinguish pathologic abnormalities from normal population variation. Accordingly, the December 2007 American College of Obstetrics and Gynecology (ACOG) Practice Bulletin on prenatal diagnosis stated, "Until there are more data available, use of CGH for routine prenatal diagnosis is not recommended."⁷⁴ Neverthe-less, in selected cases with significant ultrasound abnormalities and normal karyotype, microarray-CGH appears to be emerging as an adjunctive test for specimens obtained by amniocentesis. To this end, arrays for CGH specifically in prenatal specimens have been developed. Although further data are required before they are used routinely, such arrays are now commercially available and are beginning to be used in selected cases for prenatal diagnosis.

The prevalence of chromosomal abnormalities was estimated to be higher in twins than in singleton pregnancies when twins were discovered at genetic amniocentesis due to advanced maternal age. Similarly, the incidence of neural tube defects in a patient with a previous history of a child with a neural tube defect is higher in a twin gestation (as much as doubled) than in a singleton gestation. Before amniocentesis is performed in a multiple gestation, the possible risks, outcome, and management alternatives need to be discussed with the patient. The major issue is the possibility of discrepant results regarding cytogenetic diagnosis. If only 1 fetus is affected, the options available to the patient include abortion of both fetuses, continuation of the pregnancy, or selective feticide of the affected fetus. Feticide is associated with potential complications, such as infection, disseminated intravascular coagulation, and spontaneous abortion. Under these circumstances, pregnancy termination implies the abortion of an unaffected fetus.

The technique for amniocentesis in multiple gestations is different from that in singleton gestations. The number of fetuses, location within the uterine cavity, placentation, presence of an interamniotic membrane, and fetal sex, biometry, and anatomy need to be documented. An important step is the topographic location of the fetuses. This becomes critical in cases where discrepant results are reported and selective feticide is considered. Identification of the fetuses should be based on the relation to the maternal hemipelvis (left vs right, anterior vs posterior, and superior vs inferior). We recommend that a diagram of the procedure be drawn and kept in the medical record for reference.

Several techniques have been proposed to ensure that the amniotic fluid is retrieved from each amniotic cav-ity^{152,153,154,155, and 156} and have been the subject of review by Weisz and Rodeck.¹⁵⁷ The most common technique is the sequential insertion of 2 different needles under ultrasound guidance into each amniotic cavity. After insertion of the first needle and aspiration of amniotic fluid, the needle is removed and the second sac is tapped with a new needle. Resampling of the first sac using this technique has been reported in up to 3.5% of samples. To reduce the rate of this complication, it has been proposed that, after amniotic fluid is obtained from the first sac and before removing the needle, an indicator dye is injected into the cavity. Indigo carmine is presently the dye of choice. Other alternatives are Congo red and Evans blue. The use of a solution of maternal hemoglobin as a dye has also been reported. The use of methylene blue is discouraged because of the risk of fetal hemolytic anemia due to methemoglobinemia and a possible association with gastrointestinal obstruc-tion. The association between the use of methylene blue and multiple ileal occlusions was first reported by Nicolini and Monni¹⁵⁸ in 7 neonates born to mothers who had a mid-trimester amniocentesis and the injection of 10 to 30 mg of methylene blue into one of the amniotic sacs. Pruggmayer et al¹⁵⁹ compared the incidence of gastrointestinal obstruction among 474 twin pregnancies who had either indigo carmine (n = 351) or methylene blue (n = 123) injected as a dye during mid-trimester amniocentesis. Seventeen percent (21 of 123) of the fetuses who had the sacs injected with methylene blue required a postnatal operation to correct jejunal atresia as opposed to 0.3% (1 to 351) of fetuses who had the sacs injected with indigo carmine and who presented with duodenal atresia in the neonatal period (P < .0001). It is expected that clear amniotic fluid will be obtained when the second sac is punctured. The same procedure is applicable to amniocentesis for a multiple gestation with more than 2 fetuses. The technique consists of sequential injections of dye into different sacs before removal of the needle. The number of clear amniotic fluid aspirations should equal the number of fetuses.

A technique for single needle insertion in twin amniocentesis was described by Jeanty et al.¹⁵⁴ Briefly, a puncture site clear of placental tissue demonstrating both gestational sacs and the interamniotic membrane is selected with real-time ultrasound. The most proximal sac is tapped first and the amniotic fluid is collected, the stylet is then replaced into the needle and advanced under ultrasound guidance through the interamniotic membrane into the second sac and 1 mL of fluid is aspirated and discarded to reduce the risk of contamination with amniotic fluid from the first sac. Finally, the fluid is aspirated from the second sac. Using this technique, amniotic fluid was successfully retrieved from both sacs in 94% (17 of 18) of patients included in their study, and no adverse perinatal outcomes were reported. This technique has some limitations: (1) it may be difficult to enter the second sac due to tenting of the septum; (2) contamination of the second fluid with cells from the first sac can occur; and (3) the possible risk of rupturing the interamniotic membrane and creating a pseudo-monoamniotic twin pregnancy with its associated complication of cord entanglement.^157,160

The last technique, proposed by Bahado-Singh et al,¹⁵⁶ is the simultaneous insertion of 2 needles by 2 operators into 2 separate sacs under direct ultrasonographic guidance. The procedure has been reported in a series of 7 patients, with successful retrieval of fluid in all cases and no perinatal complications. A shortcoming of this technique is that it may not be used by a single operator. The overall success rate in obtaining fluid from both sacs is greater than 90%. A challenging situation occurs when one sac is behind the other. In this case, the second is sampled by advancing the needle to penetrate the interamniotic membrane under direct visualization, as described earlier. The first 2 cc of amniotic fluid retrieved from the second sac are discarded to decrease the likelihood of contamination.

Another difficult situation arises when an interamniotic membrane cannot be identified. This can occur in the setting of polyhydramnios and in monoamniotic twins. A practical approach consists of sampling 2 sites in close proximity to each fetus but distant from each other. We have found it helpful to place the transducer along the transverse axis of the uterus and to monitor the turbulence created by the injection of the indicator dye. Before injection, the dye is diluted with 10 cc of amniotic fluid. The mixture is then injected into the amniotic cavity, producing a typical particulate image that identifies the boundaries of that sac (Figure 25-10). The operator must be ready to proceed with the second puncture because the image is short-lived. Tabsh¹⁵⁵ described another interesting approach. After aspirating the amniotic fluid sample from the first sac, 0.1 mL of air and 0.5 mL of dye are drawn through a 15-μm filter into a 6-mL syringe. The syringe is then reattached to the needle hub, and 5 mL of amniotic fluid is aspirated into the syringe. The mixture of dye, air, and amniotic fluid is then reinjected into the amniotic sac with gentle pressure, but enough to create microbubbles that will serve as an ultrasonographic contrast within the amniotic sac. If the microbubbles are seen around both fetuses, the diagnosis of monoamniotic twinning is made.

The risk of pregnancy loss after genetic amniocentesis in multiple gestations has been addressed in several reports (Table 25-6).^{159,161,162,163,164,165,166,167,168,169,170,171,172,173,174, and 175} Anderson et al¹⁷¹ compared the risk of spontaneous abortion after mid-trimester amniocentesis in 353 twin pregnancies with 687 singleton pregnancies matched for gestational age and indication for prenatal diagnosis (one preceding and one following each twin). Fourteen cases were excluded because of congenital anomalies, growth retardation, twin-to-twin transfusion syn-drome, or death of one twin at the time of the procedure, leaving 339 patients for subsequent analysis. The groups were not different with regard to maternal age or experience of the physician performing the procedure. Three pregnancies were terminated after the diagnosis of aneuploidy in one of the fetuses. After terminations for chromosomal abnormalities were excluded, the spontaneous fetal loss rate before 28 weeks was significantly higher for multiple pregnancies than for singleton pregnancies (3.6% [12 of 336] vs 0.6% [4 of 671], P < .001). The perinatal mortality rate, however, was not significantly different between the 2 groups (12.6 of 1000 [8 of 633] vs 12.1 of 1000 [8 of 659]). In another large series¹⁵⁹ assessing the perinatal outcome of 529 twin pregnancies undergoing genetic amniocentesis, the combined fetal loss rate before 28 weeks was 3.8% (20 of 529) and comparable to the 3.6% rate reported by Anderson et al.¹⁷¹ Both studies concluded that, although there is an increased risk of spontaneous pregnancy loss for twin pregnancies undergoing amniocentesis, it is unlikely that this increased risk exceeds the normal bio-logic loss rate in twins. This conclusion is also supported by the analysis of the risks of mid-trimester amniocente-sis in twin gestations conducted by Ager and Oliver,⁷⁰ although with a different outcome measure: total fetal loss rate. Total fetal loss rate is defined as follows: total spontaneous abortions + stillbirths + neonatal deaths. According to their analysis, the total fetal loss rate is 10.8% (range 3.6% to 22.2%), which is not different from the natural fetal loss rate for twin gestations after 17 weeks as estimated from 12,392 twin pregnancies collected in Japan.¹⁷⁶

Definition

The traditional view is to consider the amniotic cavity sterile; therefore, isolation of any microorganisms from the amniotic fluid constitutes evidence of microbial invasion. Microorganisms may gain access to the amniotic cavity and fetus using any of the following pathways: (1) ascending from the vagina and the cervix, (2) hematogenous dissemination through the placenta (transplacental infection), (3) retrograde seeding from the peritoneal cavity through the fallopian tubes, and (4) accidental introduction at the time of invasive procedures, such as amniocente-sis, percutaneous fetal blood sampling, CVS, or shunting. The most common pathway of intrauterine infection is the ascending route.^{177,178, and 179}

We have proposed a 4-stage process leading to intrauterine infection (Figure 25-11 modified from Kim et al).¹⁸⁰ The first stage consists of an overgrowth of facultative organisms or the presence of the pathologic organisms (ie, Neisseria gonorrhoeae) in the vagina, cervix, or both. Bacterial vaginosis may be an early manifestation of stage I. Then, microorganisms gain access to the intrauterine cavity, stage II. followed in the amniotic proliferation leading to microbial invasion of the amniotic cavity (stage III). Rupture of the membranes is not a prerequisite for this condition, as microorganisms are capable of crossing intact membranes. Once in the amniotic cavity, the bacteria may gain access to the fetus (stage IV) by different ports of entry. Aspiration of infected fluid by the fetus may lead to congenital pneumonia. Otitis, conjunctivitis, and omphalitis are localized infections that occur by direct spread of microorganisms from infected amniotic fluid. Seeding from any of these sites to the fetal circulation may lead to bacteremia and sepsis.

Frequency

The frequency of microbial invasion of the amniotic cavity based on conventional cultivation techniques, differs according to gestational age, presence of labor, cervical dilation, and state of the fetal membranes, and it ranges from 0.4% in patients in the mid-trimester of pregnancy to 51.5% in patients with cervical incompetence (Table 25-7).^{59,179,181,182,183,184,185,186,187,188,189, and 190}

Diagnosis of Microbial Invasion of the Amniotic Cavity Using Conventional Culture Techniques

The most common microbial isolates from the amniotic cavity from women with preterm labor and intact membranes and women with preterm prelabor rupture of the membranes are U urealyticum, Fusobacterium species, and M hominis (Table 25-8).^{59,179,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196, and 197}

Microbial Invasion of the Amniotic Cavity in Preterm Labor with Intact Membranes

The mean rate of positive amniotic fluid cultures for microorganisms in patients with preterm labor and intact membranes is approximately 12.8% (379 of 2963)^198,199 based upon the review of 33 studies (Table 25-9).^{179,191,192,195,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227, and 228} Women with positive amniotic fluid cultures generally do not have clinical evidence of infection at presentation, but they are more likely to develop clinical chorioamnionitis (37.5% [60 of 160] vs 9% [27 of 301]), to be refractory to tocolysis (85.6% [101 fo 118] vs 16.3% [8 of 49]), and to spontaneously rupture their membranes (40% [6 of 15] vs 3.8% [2 of 52]) than women with negative amniotic fluid cultures. Several investigators have demonstrated that the rate of neonatal complications is higher in neonates born to women with microbial invasion of the amniotic cavity than in those without infection. Moreover, the earlier the gestational age at preterm birth, the more likely that microbial invasion of the amniotic cavity was pre-sent.²¹⁴

Microbial Invasion of the Amniotic Cavity in Patients with Preterm PROM

The rate of positive amniotic fluid cultures for microorganisms is approximately 32.4% (473 of 1462)^{181,182,183,184,185, and 186,188,211,213,220,222,225,229,230,231,232, and 234} in patients with preterm PROM (Table 25-10).¹⁹⁹ Clinical chorioamnionitis is present in only 29.7% (49 of 165) of the cases with proven microbial inva-sion.^199,235 The rate of microbial invasion in preterm PROM reported by these studies is probably an underestimation of the true prevalence of intrauterine infection. Indeed, available evidence indicates that women with preterm PROM and severely reduced volume of amniotic fluid have a higher incidence of intra-amniotic infection than those without oligohydramnios.^185,236 Since women with oligohydramnios are less likely to have an amniocentesis, the bias in these studies is to underestimate the prevalence of infection. A similar bias is that women with preterm PROM admitted in labor generally do not undergo amniocentesis. These patients have a higher rate of microbial invasion of the amniotic cavity than those admitted without labor (39% [24 of 61] vs 25% [41 of 160], P < .05). Furthermore, of patients who are not in labor on admission, 75% have a positive amniotic fluid culture at the time of the onset of labor.¹⁸⁶ Therefore, studies restricted to women not in labor provide a lower rate of microbial invasion of the amniotic cavity than those including patients in labor.

Microbial Invasion of the Amniotic Cavity in Patients with a Sonographic Short Cervix

A sonographic short cervix is considered a powerful predictor of spontaneous pre-term delivery. Indeed, the shorter the sonographic cervical length in the mid-trimester of pregnancy, the higher the likelihood of spontaneous preterm delivery. Guzman et al²³⁷ studied 189 patients (125 singleton, 45 twin, and 19 triplet pregnancies) that underwent serial transvaginal cervical measurements with transfundal pressure between 15 and 24 weeks of gestation. The results showed that patients with a sonographic cervical length less than 20 mm had a significantly higher rate of acute inflammatory lesions in placental histologic examinations than those with a cervical length 20 mm or greater (48.6% vs 26.5%, P = .003). Hassan et al²³⁸ conducted a retrospective cohort study of 152 patients with a sonographic short cervix (<25 mm) at 14 to 24 weeks of gestation. Amniocenteses were performed in 57 patients and the prevalence of intra-amniotc infection was 9% (5 of 57); among these patients, the rate of preterm delivery at less than 32 weeks was 40% (2 of 5), suggesting that a short cervix may be the only manifestation of subclinical microbial invasion of the amniotic cavity.

Microbial Invasion of the Amniotic Cavity in Patients Presenting with Acute Cervical Incompetence

Women presenting with a dilated cervix, intact membranes, and few, if any, contractions before the 24th week are considered to have clinical cervical incompetence. In one study, 51.1% of these patients had a positive amniotic fluid culture for microorganisms.¹⁸⁹ The outcome of patients with microbial invasion was uniformly poor because they developed subsequent complications (rupture of membranes, clinical chorioamnioni-tis, or pregnancy loss). Recently, Bujold et al²³⁹ conducted a study including 15 patients with suspected cervical insufficiency between 16 and 26 weeks of gestation. The diagnosis of intra-amniotic infection was confirmed in 47% (7 of 15), of whom 20% (3 of 15) were infected with more than one bacterial strain and 33% (5 of 15) with Ureaplasma species. This evidence suggests that infection is frequently associated with acute cervical incompetence. Whether intra-amniotic infection is the cause or consequence of cervical dilatation has not been determined. It is possible that clinical silent cervical dilatation with protrusion of the membranes into the vagina would lead to a secondary intrauterine infection.

Microbial Invasion of the Amniotic Cavity in Patients with Twin Gestations and Preterm Labor

The prevalence and outcome of microbial invasion of the amniotic cavity in twin gestation presenting with preterm labor and intact membranes was the subject of investigation almost 2 decades ago.¹⁸⁷ Amniocenteses were performed on both sacs of 46 women with twin gestations, preterm labor, and intact membranes. Overall, 91.3% (42 of 46) of women delivered prematurely. A positive amniotic fluid culture of at least 1 sac was noted in 10.8% (5 of 46) of patients admitted in preterm labor and in 11.9% (5 of 42) of those who delivered preterm. Among those 5 patients with microbial invasion of the amniotic cavity, 3 had microorganisms present in the amniotic fluid isolated from both sacs. The presenting sac was involved in all cases, supporting an ascending route for microbial invasion of the amniotic cavity in twin gestation. Similar findings were observed by Mazor et al,²⁴⁰ who conducted a study including 74 consecutive twin pregnancies with spontaneous preterm labor with intact membranes that underwent transabdominal amniocentesis under sonographic guidance. The prevalence of preterm delivery was 91.9% (68 of 74). Amniotic fluid was retrieved from both sacs, and culture results were positive for microorganisms in 12.1% (9 of 74). Among patients with intra-amniotic infection, microorganisms were isolated from the presenting sac in 5 cases (55.6%), from both sacs in 3 patients (33.3%), and from the upper sac in the remaining case (11.1%). All patients with intra-amniotic infection delivered prematurely and had a more advanced cervical dilatation, a shorter interval from amniocentesis to de-livery, and a higher incidence of clinical chorioamnionitis than those with a negative amniotic fluid culture. These findings are in contrast to the 21.6% of abnormal amniotic fluid cultures observed in singleton gestations with pre-term labor and delivery,¹⁷⁹ suggesting that intra-amniotic infection is a possible cause of preterm labor and delivery in twin gestation, but they do not support the hypothesis that intra-amniotic infection is responsible for the excessive rate of preterm delivery observed in these patients.

Diagnosis of Microbial Invasion of the Amniotic Cavity Using Molecular Microbiology Techniques

The prevalence of microbial invasion of the amniotic cavity previously described is based on the results of standard microbiologic methods. A positive culture in amniotic fluid can only be obtained if the culture conditions in the laboratory are able to support the growth of a particular microorganism. Since the growth requirements of all microorganisms are unknown, a negative culture should not be taken to definitively exclude the presence of microorganism. In other words, while a positive culture is indicative of microbial invasion of the amniotic cavity, a negative culture indicates that the laboratory was not able to grow bacteria from the specimen, either because bacteria was absent (a true negative result) or because the laboratory conditions did not support the growth of a specific microorganism (a false-negative result). Interestingly, only 1% of the whole microbial world can be detected by cultivation techniques ("the great plate count anomaly"). Thus, the frequency of microbial invasion of the amniotic cavity reported in the literature represents minimum estimates that are likely to change with the introduction of more sensitive methods for microbial recovery and identification. Indeed, several investigators have demonstrated that the prevalence of mi-crobial invasion of the amniotic cavity is higher when molecular microbiologic techniques are used to detect conserved sequences in prokaryotes (eg, bacterial 16S rDNA with PCR). Two strategies have been used to detect mi-croorganisms in the amniotic fluid using PCR. The first, also known as broad-range PCR, uses primer pairs designed to anneal with highly conserved DNA regions of all bacteria. A positive result indicates the presence of bacteria; however, identification of the specific microorganism requires subsequent sequencing of the PCR products. The second approach is to use specific primers for a particular microorganism.

Blanchard et al²⁴¹ were the first to report the recovery of U urealyticum using specific primers for the urease structural genes from 293 amniotic fluid samples collected by amniocentesis at cesarean section. The samples were cultured for bacteria, mycoplasmas, and chlamydiae, and had PCR performed for U urealyticum. Among the 10 PCR-positive amniotic fluid samples, only 4 were also culture positive. Three subsequent studies applied broad-spectrum bacterial 16S rDNA PCR for the detection of bacteria in the amniotic fluid. Jalava et al²⁴² studied 20 amniotic fluid samples from patients with PROM and 16 control samples: PCR detected 5 microorganisms (U urealyticum [n = 2], Haemophilus influenzae [n = 1], Streptococcus oralis [n = 1], and Fusobacterium species [n=1]), whereas only 2 samples were positive in a traditional bacterial cul-ture. Two patients developed infectious complications, and both were correctly identified by PCR. Hitti et al²⁴³ studied 69 women in preterm labor with intact membranes. PCR was positive in 94% of the culture-positive amniotic fluid samples (15 of 16), and 36% (5 of 14) of patients with a negative culture had an elevated IL-6 concentration. Markenson et al²¹⁸ examined amniotic fluid samples from 54 women in preterm labor without clinical evidence of intra-amniotic infection. Interestingly, 55.5% of amniotic fluid samples were PCR positive, whereas only 9.2% of the cultures recovered a microorganism (P < .05).

Oyarzun et al²²⁴ described a PCR amplification technique aimed at detecting the 16 microor-ganisms most commonly cultured from the amniotic fluid of patients in preterm labor (U urealyticum, M hominis, Gardnerella vaginalis, E coli, Fusobacterium species, Peptostreptococcus anaerobius, Bacteroides fragilis, Chlamydia trachomatis, H influenzae, N gonorrhoeae, and Streptococcus species). Amniotic fluid samples were examined with bacterial culture and PCR in 50 patients with preterm labor and intact membranes and 23 patients not in labor. All control samples were both culture and PCR negative. A higher proportion of samples were positive by PCR compared to culture (46% [23 of 50] vs 12% [6 of 50], P < .05). However, there were patients with positive PCR who delivered at term without maternal or neonatal complications.

The question of the clinical significance of microbial invasion of the amniotic cavity detected purely by molecular microbiology techniques, but which cannot be detected by cultivation techniques, was evaluated by Yoon et al in patients with preterm PROM²³⁴ or preterm labor and intact membranes.²⁴⁴ Patients with a positive PCR assay of amniotic fluid, but a negative culture, had a stronger amniotic fluid inflammatory reaction (higher amniotic fluid white blood cell count and IL-6 concentrations), a shorter interval to delivery, as well as a higher rate of histologic chorioamnionitis, funisitis, and neonatal morbid events than those with a negative amniotic fluid culture and negative amniotic fluid PCR assay for U urealyticum. These studies suggest that patients with preterm PROM or preterm labor with intact membranes and a negative amniotic fluid culture, but a positive PCR assay for U urealyticum, have worse pregnancy outcomes than those with a sterile amniotic fluid culture and negative PCR.

DiGiulio et al used broad-range end-point and real-time PCR assays to amplify, identify, and quantify ribosomal DNA (rDNA) of bacteria, fungi, and archaea from amniotic fluid of 166 consecutive women with spontaneous preterm labor with intact membranes.²⁴⁶ The combined use of molecular and culture methods revealed a greater prevalence (15%) and diversity (18 taxa) of microbes in amniotic fluid than did culture alone (9.6%; 11 taxa). A positive PCR was associated with histologic chorioamnionitis (adjusted odds ratio [OR] 20; 95% CI, 2.4 to 172), and funisitis (adjusted OR 18; 95% CI, 3.1 to 99). The positive predictive value of PCR for preterm delivery was 100%. A temporal association between a positive PCR and delivery was supported by a shortened amniocentesis-to-delivery interval. This study shows that the amniotic cavity of women in preterm labor harbors DNA from a greater diversity of microbes than previously suspected, including as-yet uncultivated, previously-uncharacterized taxa. A dose-response association was demonstrated between bacterial rDNA abundance and gestational age at delivery. The strength, temporality, and gradient with which these microbial sequence types are associated with preterm delivery support a causal relationship.²⁴⁶ We also analyzed the amniotic fluid of 204 consecutive women with preterm PROM using both cultivation and molecular methods.²⁴⁷ The prevalence of microbial invasion of the amniotic cavity was 34% (70/204) by culture, 45% (92/204) by PCR, and 50% (101/204) by both methods combined. The number of bacterial species revealed by PCR (44 species-level phylotypes) was greater than that by culture (14 species) and included as-yet uncultivated taxa. Some taxa detected by PCR have been previously associated with the gastrointestinal tract (e.g., Coprobacillus sp.), the mouth (e.g., Rothia dentocariosa) or the vagina in the setting of bacterial vaginosis (e.g., Atopobium vaginae). Bacterial rDNA abundance exhibited a dose relationship with gestational age at delivery. A positive PCR was associated with lower mean birthweight, and with higher rates of respiratory distress syndrome and necrotizing enterocolitis. These results suggest that MIAC in preterm PROM is more common than previously recognized and is associated in some cases with uncultivated taxa, some of which are typically associated with the gastrointestinal tract.²⁴⁷

Gerber et al²⁴⁸ identified U urealyticum using PCR in 11% (29 of 254) of amniotic fluid samples obtained between 15 and 17 weeks of gestation. Patients that were positive for U urealyti-cum had a significantly higher rate of subsequent preterm labor (58% [17 of 29] vs 4% [10 of 225]; P < .0001) and preterm birth (24% [7 of 29] vs 0.4% [1 of 225]; P < .0001) than those who had a negative result. In a smaller study²⁴⁹ of 78 patients undergoing genetic amniocentesis, bacterial 16S ribosomal DNA was detected in 18% (14 of 78) of the amniotic fluid samples. In this study, no samples tested positive for Mycoplasma species DNA, although it was specifically sought using a PCR/enzyme-linked immunosorbent assay. In contrast to previous and subsequent results, no association was found between the recovery of bacteria at the time of amniocentesis and either increased IL-6 concentration in the amniotic fluid or pre-term delivery. Nguyen et al²⁵⁰ identified M hominis in 6.4% (29 of 456) of amniotic fluid samples obtained between 15 and 17 weeks of gestation from 456 women. The rate of preterm labor in women with a positive PCR for M hominis was higher than in those with negative PCR (14% vs 3%; P = .01). Finally, Perni et al²⁵¹ performed genetic amniocentesis in 179 asymptomatic women between 15 and 19 weeks of gestation. M hominis and U urealyticum were detected by PCR in 6.1% and 12.8% of cases, respectively. Preterm PROM occurred in 2.8% (5 of 179) of women, and the same proportion of patients had a spontaneous preterm birth with intact membranes. All women with preterm PROM were positive for either M hominis or U urealyticum, whereas none of those with spontaneous pre-term birth were. These observations suggest that microbial invasion of the amniotic cavity could be subclinical in the mid-trimester of pregnancy and chronic in nature, and that pregnancy loss or preterm delivery due to infection may take weeks to occur.

Importance of the Diagnosis

Accumulating evidence supports an association between microbial invasion of the amniotic cavity and preterm labor and delivery.¹⁹⁰ Microbial invasion of the amniotic cavity has been reported in 21.6% of women with spontaneous preterm labor and intact membranes who subsequently delivered a preterm neo-nate.¹⁷⁹ This condition is often subclinical in nature and requires microbiologic studies of amniotic fluid for diagnosis. The early identification of microbial invasion of the amniotic cavity is a desirable clinical goal because neonates born to mothers with intra-amniotic infection are at higher risk for both infectious and noninfectious complications (Table 25-11).^{179,181,182,183, and 184,195,201,210,215,229,230} More-over, patients with microbial invasion of the amniotic cavity are at risk for the development of clinical chorioamnio-nitis, failure to respond to tocolysis, subsequent rupture of membranes, and puerperal endometritis (see Table 25-11).¹⁹⁰ Despite the importance of microbial invasion of the amniotic cavity, an early diagnosis is difficult to accomplish. Clinical signs of infection (fever, maternal and/or fetal tachycardia, uterine tenderness, foul-smelling vaginal discharge, and maternal leukocytosis) occur late and are found only in 12% of these pa-tients.¹⁷⁹ However, amniotic fluid cultures, considered the gold standard for diagnosis, are not immediately available for clinical management, and their results may take several days. Consequently, there has been considerable interest in the development of rapid tests for the diagnosis of this condition by analysis of amniotic fluid.

Amniotic Fluid Collection

The method of amniotic fluid collection for microbiologic studies is critical. The 2 techniques that have been used are transabdominal amniocentesis and transcervical retrieval either by needle puncturing of the membranes or by aspiration through an intrauterine catheter. Transcervical amniotic fluid collection is associated with an unacceptable risk of contamination with vaginal flora and is contraindicated in both spontaneous preterm labor as well as nonlaboring patients with preterm PROM. Successful retrieval of amniotic fluid by ultrasonographically monitored am-niocentesis is possible in virtually all patients with preterm labor with intact membranes, and in more than 90% of patients with PROM, making it the method of choice to obtain amniotic fluid from these patients.

Tests for Diagnosis of Microbial Invasion of the Amniotic Cavity

In the following section, we review the clinical value of amniotic fluid Gram stain, acridine orange tests, limulus amebocyte lysate assays, white blood cell count, as well as glucose, IL-6, and matrix metalloproteinase-8 (MMP-8) concentrations in the identification of microbial invasion of the amniotic cavity.

Amniotic Fluid Gram Stain

The Gram stain is the most frequently used rapid diagnostic test for detecting intra-amniotic infection and has been used extensively for making clinical decisions in patients at risk for intra-amniotic infection. Some technical aspects should be taken into account to obtain optimal results using this test: (1) the slide should be prepared with fluid obtained directly from the syringe because swabs absorb both fluid and cells, decreasing the likelihood of observing organisms in a smear or culture; and (2) although centrifugation of amniotic fluid at low speed does not improve the detection of bacteria with the Gram stain,²⁰⁸ the use of a cytocentrifuge increases the concentration of bacteria in the sediment and probably allows an easier identification of microorganisms.

Table 25-12 shows the diagnostic accuracy of the Gram stain as a diagnostic tool in the detection of microbial invasion of the amniotic cavity in different studies.^{179,181,182,183, and 184,186,188,192,195,200,201,203,206,210,211,213,215,229,230,252,253} The sensitivity increases significantly with a higher inoculum (>10⁵ colony-forming units per milliliter).²⁰⁸ It is important to note that Mycoplasma species, which are frequently isolated in the amniotic fluid of patients in preterm labor with intact membranes or preterm PROM, are not visible on Gram stain examination.

Acridine Orange Stain

Acridine orange stain is a fluorochrome dye that, when buffered at low pH, binds to the nucleic acid of bacteria and stains them an orange color. The nucleus of human epithelial and inflammatory cells shows a green to yellow fluorescence, with no orange or green fluorescence in the cytoplasm. Therefore, bacteria should be readily identified as bright orange structures against a dark background. Several investigators have claimed that acridine orange stain is more sensitive than the Gram stain in the detection of bacteria in biologic fluids. Data generated by our group, how-ever, indicate that acridine orange stain cannot replace the Gram stain as a method to diagnose microbial invasion of the amniotic fluid, because the sensitivity, specificity, and predictive values of the test are not superior to those obtained when using Gram stain (Table 25-13).²⁵⁴ The potential use of acridine orange stain would be to identify mycoplasmas, but this potential advantage needs to be weighed against the risk of increasing the false-positive diagnosis.

Limulus Amebocyte Lysate Assay

Limulus amebocyte lysate assay is a rapid, inexpensive, and sensitive bioassay for the detection of bacterial lipopolysac-charide. It is based on the gelation of the lysate of blood cells (amebocyte) of the horseshoe crab, Limulus polyphemus, in the presence of endotoxin. Gram-negative organisms are frequently found in microbial invasion of the amniotic cavity. Their cell walls contain lipopolysaccharide, which can be detected either free after cell lysis or as part of the intact bacteria.

Gram stain and limulus amebocyte lysate have comparable diagnostic performances for the diagnosis of microbial invasion of the amniotic cavity (sensitivity of 69%, specificity of 95%), but when used in combination, there is a remarkable improvement of the sensitivity (from 69% to 96%) with a small decrease in the specificity (from 95% to 86%).²⁵⁵ Although the limulus amebocyte lysate assay test would be expected to detect only gram-negative microorganisms, it also identifies 33% of pure gram-positive infections and cases of intra-amniotic infection produced by mycoplasmas or Candida. Two explanations have been proposed for these findings. The infection could have included gram-negative organisms that were not grown in culture because of their fastidious nature or because they were not visible. In the latter case, free lipopolysaccharide released at the time of cell death could lead to a positive limulus amebocyte lysate result, although the organism may not grow in culture. An alternative explanation is that cross-reacting microbial products could induce gelation of the limulus lysate.²⁵⁵ For example, peptidoglycans, a component of gram-positive bacteria, and mannan, a carbohydrate molecule isolated from Candida species, have been reported to produce a positive limulus amebocyte lysate test result.

Glucose Concentration in Amniotic Fluid

Glucose determination has been used extensively to diagnose infection in other body fluids (ie, cerebrospinal fluid, pleural fluid, and synovial fluid). It does not require sophisticated interpretation by trained personnel. Amniotic fluid glucose concentration is significantly lower in patients with microbial invasion of the amniotic cavity (identified by a positive amniotic fluid culture or clinical signs of infection) and in patients with preterm PROM who develop clinical infection.^{210,211,213,215,230,256} The mechanism responsible for the lower amniotic fluid glucose concentration in the setting of infection has not been established, but probably relates to glucose metabolism by both microorganisms and polymorphonuclear leukocytes.²¹³

Table 25-14 displays the diagnostic indices of amniotic fluid glucose concentration in the diagnosis of microbial invasion of the amniotic cavity according to various published reports. Collectively, these data indicate that the sensitivity and specificity of amniotic fluid glucose concentrations for the diagnosis of intra-amniotic infection range from 57% to 87% and from 51% to 100%, respectively, when using different cutoff values. As in the case of white blood cell count, the false-positive rates of amniotic fluid glucose concentration are high (8% to 48%) (see Table 25-14, and 25-15). The importance of a false-positive result is dependent on the action taken after an abnormal test result. If the course of action is to perform additional tests for the identification of intra-amniotic infection, then a false-positive result is not clinically problematic; however, if the course of action is to deliver a preterm neonate believed to be in-fected, then there is the potential for serious consequences.²¹⁰

White Blood Cell Count in Amniotic Fluid

Neutrophils are infrequently found in the amniotic fluid of women who are not in labor. Their presence in the amniotic fluid indicates the existence of an inflammatory reaction, usually caused by microbial invasion of the amniotic cavity. Table 25-15 summarizes the results of 3 different studies examining the performance of amniotic fluid white blood cell count in the identification of microbial invasion of the amniotic cavity in patients with spontaneous preterm labor and intact membranes and in women with preterm PROM.^212,215,230 In a study of 195 patients in preterm labor with intact membranes who underwent amniocentesis for the assessment of the microbiologic status of the amniotic cavity, a white blood cell count greater than or equal to 50 cells/mm³ had a sensitivity of 80% (20 of 25) and a specificity of 87.6% (149 of 170) in the detection of a positive amniotic fluid culture.²¹² Although the sensitivity was higher than the Gram stain (80% vs 48%, P < .05), the specificity was lower (87.6% vs 98.8%, P < .05); thus, the false-positive rate was high (12.4%). Interestingly, patients with an amniotic fluid white blood cell count greater than or equal to 50 cells/mm³ but a negative amniotic fluid culture are at risk for preterm delivery.²¹² Therefore, independently of the culture results, an elevated amniotic fluid white blood cell count identifies a subset of patients at risk for failure to respond to tocolysis and impending preterm delivery. These patients may have microbiologic invasion of the amniotic cavity that escapes detection with standard microbiologic techniques, or a noninfectious process may have driven neutrophil recruitment into the amniotic cavity. We have proposed a different cutoff value of white blood cell count for the diagnosis of microbial invasion of the amniotic cavity in patients with preterm PROM (30 cells/mm³) (Figure 25-12).²³⁰

Amniotic Fluid Interleukin-6 Concentrations

Microbial invasion of the amniotic cavity, elevated pro-inflammatory cytokines in the amniotic fluid and histologic chorioamnionitis are part of the "inflammatory cluster" related to patients with preterm parturition.²⁵⁴ Intra-amniotic infection and/or inflammation is present in approximately one-third of the patients with spontaneous preterm labor,^199,258,259 is associated with development of fetal inflammatory response syn-drome,^220,260 and is considered a risk factor for fetal injury.^{261,262,263,264,265,266,267,268,269,270,271, and 272}

Interleukin-6 has been implicated as a major mediator of the host response to infection and tissue damage. A growing body of evidence indicates that preterm parturition in the setting of infection is associated with dramatic alterations in the amniotic fluid concentration of several cytokines, including IL-6.^{209,274,275,276,277, and 278} This cytokine is a glycoprotein that is produced in a wide variety of cells, such as fibroblasts, monocyte/macrophages, endothelial cells, keratinocytes, and endometrial stromal cells. Interleukin-6 expression is induced by several in-flammation-associated cytokines, including IL-1, tumor necrosis factor and interferons, bacterial products, RNA- and DNA-containing viruses, and second messenger agonists (diacylglycerol, cAMP, and Ca²⁺) that activate any of the 3 major signal transduction pathways. Interleukin-6 elicits major changes in the biochemi-cal, physiologic, and immunologic status of the host, including the acute phase plasma protein response, activation of T and natural killer cells, and stimulation and proliferation of immunoglobulins production by B cells. Inter-leukin-6 also induces the production of C-reactive protein (CRP). This may be important in the context of intra-amniotic infection, because clinical studies have indicated that an increase in maternal serum CRP often pre-cedes the development of clinical chorioamnionitis and the onset of preterm labor in women with preterm PROM. In addition, it has been demonstrated that IL-6 stimulates prostaglandin production by human amnion and decidual cells in primary cultures.²⁸⁰

Interleukin-6 has been studied as a rapid test for the detection of microbial invasion of the amniotic cav-ity.^211,215,230 Patients with a positive amniotic fluid culture have significantly higher amniotic fluid IL-6 concentrations than patients with a negative culture. In patients with preterm labor and intact membranes, a cutoff level of 11.3 ng/mL has a sensitivity of 100% and a specificity of 82.6% in the detection of microbial invasion of the amniotic cavity (Table 25-16).²⁷⁹ Among patients with PROM, a cutoff level of 7.9 ng/mL results in a sensitivity of 80.9% and a specificity of 75% (Table 25-17).²³⁰ Moreover, an elevated amniotic fluid IL-6 is an independent predictor of the likelihood of preterm delivery and a risk factor for neonatal complications.^215,230

Another important clinical role for IL-6 determination has been proposed by our group in patients undergoing mid-trimester amniocentesis.²⁸¹ Amniotic fluid IL-6 concentrations were higher in patients with a pregnancy loss than in patients with normal pregnancy outcome. An amniotic fluid IL-6 concentration greater than or equal to 2.8 ng/mL has been associated with an odds ratio of 8.1 (95% CI, 1.9 to 36.3) for pregnancy loss. We propose that a preexisting but subclinical inflammatory process is an important risk factor for pregnancy loss after mid-trimester amniocentesis.²⁸¹ This finding may have relevant clinical and legal implications.

An interesting observation is that the outcome of patients with spontaneous preterm labor with intra-amniotic inflammation and a negative amniotic fluid culture is similar to that of patients with microbiologically proven intra-amniotic infection. Yoon et al²⁵⁹ performed amniocentesis in 206 patients with spontaneous pre-term labor and intact membranes. Amniotic fluid was cultured for aerobic and anaerobic bacteria and mycoplas-mas, and the diagnosis of intra-amniotic inflammation was made on the basis of amniotic fluid concentrations of IL-6 greater than or equal to 2.6 ng/mL, a cutoff derived from a receiver operating characteristic (ROC) curve. Statistical analysis was conducted with contingency tables and survival techniques. Intra-amniotic inflammation (negative amniotic fluid culture but elevated amniotic IL-6) was more common than intra-amniotic infection (posi-tive amniotic fluid culture regardless of amniotic fluid IL-6 concentration; 21% [44 of 206 women] vs 10% [21 of 206 women]; P < .001]. Spontaneous preterm delivery at less than 37 weeks was more frequent in patients with intra-amniotic inflammation than in those with a negative culture and without inflammation (98% vs 35%; P < .001). Moreover, patients with intra-amniotic inflammation had a significantly higher rate of adverse outcome than patients with a negative culture and without intra-amniotic inflammation, including clinical and histologic chorioamnionitis, funisitis, early preterm birth, and significant neonatal morbidity. There were no significant differences in the rate of adverse outcomes between patients with a negative culture but with intra-amniotic inflammation and patients with intra-amniotic infection. The authors concluded that the outcome of patients with microbiologically proven intra-amniotic infection is similar to that of patients with intra-amniotic inflammation and a negative amniotic fluid culture. Similar observations have been reported in patients with preterm PROM.²⁸²

Matrix Metalloproteinase-8 in Amniotic Fluid

Matrix metalloproteinase-8, or human neutrophil collagenase, is a glycoprotein that is synthesized as a latent pro-enzyme during the myelocyte stage of neutrophil development and is released during inflammatory conditions. Matrix metalloproteinase-8 is contained in the secondary or specific granules of polymorphonuclear leukocytes and released after stimulation of polymorphonuclear leukocytes by proinflammatory stimuli. The active form of this enzyme has been implicated in the cleavage of collagen types I, II, and III, as well as in proteolytic damage to connective tissues in inflammatory conditions. Matrix metalloproteinase-8 plays an important role in microbial invasion of the amniotic cavity, preterm PROM, and term and preterm labor. It has been reported that spontaneous labor, both term and preterm, is associated with a significant increase in amniotic fluid concentrations of MMP-8.²⁸³ Similarly, microbial invasion of the amniotic cavity is associated with a significant increase in amniotic fluid MMP-8 concentration in patients with preterm labor and intact membranes,²⁸¹ as well as in patients with preterm PROM.^279,282 Indeed, among patients with spontaneous preterm labor, the median amniotic fluid concentration of MMP-8 is more than 50-fold higher in patients with intra-amniotic infection than in patients without, and is significantly higher in patients with preterm labor who delivered preterm than in those who delivered at term, even after the exclusion of patients with intra-amniotic infection.²⁸⁴

Comparison between Amniotic Fluid Tests

The diagnostic performance of amniotic fluid IL-6 determination, glucose concentration, white blood cell count, and Gram stain in patients with preterm labor and intact membranes²¹⁵ and patients with PROM²³⁰ has been studied. Figure 25-13 shows the ROC curves for the performance of amniotic fluid IL-6, white blood cell count, and glucose in the detection of a positive amniotic fluid culture in patients with preterm labor and intact membranes.²¹⁵ Diagnostic indices and prognostic values for each test and their potential combinations are displayed in Table 25-16. Interleukin-6 was the most sensitive test for the detection of microbial invasion of the amniotic cavity (100%), followed by glucose concentration (81.8%), white blood cell count (63.6%), and Gram stain (63.6%). The most specific test was Gram stain (99.1%), followed by white blood cell count (94.5%), IL-6 (82.6%), and glucose (81.6%).

Figure 25-12 shows the ROC curves for amniotic fluid IL-6, white blood cell count, and glucose concentration in the detection of a positive amniotic fluid culture in patients with preterm PROM.²³⁰ Interleukin-6 was the most sensitive individual test in the detection of microbial invasion of the amniotic cavity (80.9%), followed by MMP-8 (76.1%), glucose concentration (71.4%), and Gram stain (34.8%). The most specific test for the detection of microbial invasion of the amniotic cavity was the Gram stain (96.4%), followed by white blood cell count (83.8%), IL-6 (75%), and glucose concentration (73.5%) (see Table 25-17).²³⁰ Interleukin-6 was the only proven test to have an independent relationship with the amniocentesis to delivery interval and likelihood of neonatal complications.

Bedside Rapid Test for Diagnosis of Microbial Invasion of the Amniotic Cavity

Significant progress has been made in the development of rapid tests for the diagnosis of microbial invasion of the amniotic cavity. Although Gram stain, white blood cell count, and glucose determination remain valuable tools for clinical decision making, and the determination of IL-6 in amniotic fluid seems to be better than any other current method for the detection of microbial invasion of the amniotic cavity, MMP-8 has demonstrated to be an excellent marker of intra-amniotic infection. A comparison between the ROC curves demonstrated that the diagnostic performance of amniotic fluid MMP-8 concentration is superior to that of IL-6 and amniotic fluid white blood cell count for the prediction of a positive amniotic fluid culture (P < .001).²⁸⁴

Since the outcome of patients with microbiologically proven intra-amniotic infection is similar to that of patients with intra-amniotic inflammation and a negative amniotic fluid culture, our group has proposed that the management of patients with spontaneous preterm labor²⁵⁹ or preterm PROM²⁷⁹ should be based on the operational diagnosis of intra-amniotic inflammation rather than the diagnosis of intra-amniotic infection, because the results of the amniotic fluid culture will be available not earlier than 48 hours. Recently, a bedside test was developed to detect intra-amniotic inflammation that is based on the detection of an elevated concentration of MMP-8 in amniotic fluid.²⁸⁶ The configuration of the MMP-8 PTD (preterm deliv-ery) Check test (SK Pharma Co, Ltd, Kyunggi-do, Korea) is similar to a rapid pregnancy test, and some of its advantages are that it requires only 20 μL of amniotic fluid and no laboratory equipment, and that the results are available within 15 minutes. Our group conducted a prospective cohort study²⁸⁷ including 331 patients admitted with increased uterine contractility and intact membranes between 22 and 35 weeks of gesta-tion. The prevalence of intra-amniotic infection, intra-amniotic inflammation, and preterm delivery at less than 37 weeks of gestation were 7.3% (24 of 331), 11.5% (38 of 331), and 41.1% (136 of 331), respectively. Table 25-18 displays the diagnostic indices, predictive values, and likelihood ratios of the MMP-8 rapid test for the detection of intra-amniotic infection and inflammation. The efficiency of a positive MMP-8 rapid test result in the identification of intra-amniotic infection and intra-amniotic inflammation was 94% (311 of 331) and 97% (321 of 331), re-spectively. Table 25-19 shows the diagnostic indices, predictive values, and likelihood ratios of the MMP-8 rapid test for the identification of patients with spontaneous preterm delivery within 48 hours, 7 days, 14 days, and at less than 32 and less than 34 weeks of gestation. Moreover, patients with a positive MMP-8 rapid test result had a significantly shorter amniocentesis-to-delivery interval than patients with a negative MMP-8 rapid test result (mean, 3 days [95% CI, 1 to 4 days] vs 41 days [95% CI, 38 to 44 days]; log rank, P < .001). These results demonstrated that the MMP-8 PTD Check is a sensitive and specific test for the identification of both intra-amniotic infection and inflammation among patients with preterm labor and intact membranes. Inter-estingly, a patient with a positive MMP-8 rapid test result is at a substantial risk for spontaneous preterm delivery within 48 hours, 7 days, and 14 days, with likelihood ratios of a positive test ranging from 17 to 61.²⁸⁷ Further clinical trials are needed to determine whether, based on the MMP-8 rapid test results, treatment with antibiotics and/or anti-inflammatory agents may improve pregnancy outcome.

Acknowledgments

This work was conducted by members of the Perinatology Research Branch of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health (NIH), and funded by the Intramural Program of NICHD/NIH.

This chapter was published in previous editions of this book. The current version has been substantially changed. The authors wish to acknowledge the contribution of authors of previous chapters. In particular, the current authors and the editors gratefully acknowledge the intellectual contributions of Dr Eli Maymon.