Articles on Medical Diseases and Conditions

Category: Tests in Obstetrics

  • Biochemical Tests for Congenital Anomalies

    Besides giving information on fetal well-being, amniocentesis makes it possible to test for various congenital anomalies via biochemical analysis of amniotic fluid and tissue culture chromosome studies of fetal cells (see Chapter 34). In addition, certain substances of fetal origin may appear in maternal serum. In some cases it is possible to detect certain fetal malformations by screening tests in maternal serum.

    Maternal serum alpha-fetoprotein

    One of the most widely publicized tests for congenital anomalies is the alpha-fetoprotein (AFP) test in maternal serum for detection of open neural tube defects. Although neural tube defects are much more common in infants born to families in which a previous child had such an abnormality, about 90% occur in families with no previous history of malformation. AFP is an alpha-1 glycoprotein with a molecular weight of about 70,000. It is first produced by the fetal yolk sac and then mostly by the fetal liver. It becomes the predominant fetal serum protein by the 12th or 13th week of gestation but then declines to about 1% of peak levels by delivery. It is excreted via fetal urine into amniotic fluid and from there reaches maternal blood. After the 13th week both fetal serum and amniotic fluid AFP levels decline in parallel, the fetal blood level being about 200 times the amniotic fluid level. In contrast, maternal serum levels become detectable at about the 12th to 14th week and reach a peak between the 26th and 32nd week of gestation. Although maternal serum screening could be done between the 15th and 20th weeks, the majority of investigators have decided that the interval between the 16th and 18th weeks is optimal, since the amniotic fluid AFP level is still relatively high and the fetus is still relatively early in gestation. Maternal AFP normal levels differ for each week of gestation and ideally should be determined for each laboratory. Results are reported as multiples (e.g., 1.5Ч, 2.3Ч) of the normal population mean value for gestational age. In any patients with abnormal AFP values it is essential to confirm fetal gestational age by ultrasound, since 50% of abnormal AFP results are found to be normal due to ultrasound findings that result in a change being made in a previously estimated gestational date. Some reports suggest that maternal weight is also a factor, with heavier women tending to have lower serum AFP values (one group of investigators does not agree that maternal AFP values should be corrected for maternal weight). There are also some reports that AFP values are affected by race, at least when comparing values from Europeans and African Americans.

    Maternal AFP levels reportedly detect about 85%-90% (literature range, 67%-97%) of open neural tube defects; about one half are anencephaly and about one half are open or closed spinabifida. There is an incidence of about 1-2 per 1,000 live births. The test also detects a lesser (but currently unknown) percentage of certain other abnormalities, such as fetal ventral wall defects, Turner’s syndrome, pilonidal sinus, hydrocephalus, duodenal atresia, multiple hypospadias, congenital nephrosis, and cystic hygroma. In addition, some cases of recent fetal death, threatened abortion, and Rh erythroblastosis produce elevated maternal AFP levels, as well as some cases of maternal chronic liver disease and some maternal serum specimens obtained soon after amniocentesis. A theoretical but unlikely consideration is AFP-producing tumors such as hepatoma. More important, twin pregnancies cause maternal values that are elevated in terms of the reference range established on single-fetus pregnancies. A large minority of elevated maternal AFP levels represent artifact due to incorrect estimation of fetal gestational age, which, in turn, would result in comparing maternal values to the wrong reference range. There is also the possibility of laboratory error. Most authorities recommend a repeat serum AFP test 1 week later to confirm an abnormal result. If results of the second specimen are abnormal, ultrasound is usually suggested to date the age of gestation more accurately, to examine the fetus for anencephaly, and to exclude twin pregnancy. However, even ultrasonic measurements may vary from true gestational age by as much as 5-7 days. Some perform ultrasonic evaluation if the first AFP test result is abnormal; if ultrasound confirms fetal abnormality, a second AFP specimen would be unnecessary. In some medical centers, about 40%-59% of elevated maternal AFP levels can be explained on the basis of technical error, incorrect fetal gestation date, and multiple pregnancy.

    Some conditions produce abnormal decrease in maternal serum AFP values. The most important is Down’s syndrome (discussed later). Other conditions that are associated with decreased maternal serum AFP levels include overestimation of fetal age and absence of pregnancy (including missed abortion).

    Amniotic fluid alpha-fetoprotein

    Amniocentesis is another technique that can be used to detect open neural tube defects. It is generally considered the next step after elevated maternal AFP levels are detected and confirmed and the age of gestation is accurately determined. As mentioned previously, amniocentesis for this purpose is generally considered to be optimal at 16-18 weeks of gestation. Assay of amniotic fluid AFP is said to be about 95% sensitive for open neural tube defects (literature range, 80%-98%), with a false positive rate in specialized centers less than 1%. Most false positive results are due to contamination by fetal blood, so a test for fetal red blood cells or hemoglobin is recommended when the amniotic fluid AFP level is elevated. Amniotic fluid AFP normal values are age related, similar to maternal serum values.

    Screening for Down’s syndrome

    While maternal serum AFP screening was being done to detect neural tube defects, it was noticed that decreased AFP levels appeared to be associated with Down’s syndrome (trisomy 21, the most common multiple malformation congenital syndrome). Previously, it had been established that women over age 35 had a higher incidence of Down’s syndrome pregnancies. In fact, although these women represent only 5%-8% of pregnancies, they account for 20%-25% (range, 14%-30%) of congenital Down’s syndrome. Since it was discovered that mothers carrying a Down’s syndrome fetus had AFP values averaging 25% below average values in normal pregnancy, it became possible to detect about 20% of all Down’s syndrome fetuses in pregnant women less than age 35 years in the second trimester. Combined with approximately 20% of all Down’s syndrome fetuses detected by amniocentesis on all possible women over age 35, the addition of AFP screening to maternal age criteria potentially detected about 40% of all Down’s syndrome pregnancies. Later, it was found that serum unconjugated estriol (uE3) was decreased about 25% below average values seen in normal pregnancies, and hCG values were increased at least 200% above average normal levels; both were independent of maternal age. Addition of hCG and uE3 to AFP screening raised the total detection rate of all Down’s syndrome patients to about 60%. Later, there was controversy whether including uE3 was cost effective. Even more recently it was found that substituting beta-hCG for total hCG increased the total Down’s syndrome detection rate to 80%-86%. Also, it was found that screening could be done in the first trimester as well as the second trimester (although AFP was less often abnormal). Finally, it was found that if AFP, uE3, and beta-hCG were all three decreased (beta-hCG decreased rather than elevated), about 60% of fetal trisomy 18 could be detected. Trisomy 18 (Edward’s syndrome) is the second most common congenital trisomy. Decreased AFP can also be caused by hydatidiform mole, insulin-dependent diabetes, and incorrect gestational age estimation.

    Amniotic fluid acetylcholinesterase

    Acetylcholinesterase (ACE) assay in amniotic fluid has been advocated as another way to detect open neural tube defects and to help eliminate diagnostic errors caused by false positive AFP results. Acetylcholinesterase is a major enzyme in spinal fluid. Results from a limited number of studies in the late 1970s and early 1980s suggest that the test has 98%-99% sensitivity for open neural tube defects. Acetylcholinesterase assay has the further advantage that it is not as dependent as AFP on gestational age. It is not specific for open neural tube defects; amniotic fluid elevations have been reported in some patients with exomphalos (protrusion of viscera outside the body due to a ventral wall defect) and certain other serious congenital anomalies and in some patients who eventually miscarry. Not enough data are available to properly evaluate risk of abnormal ACE results in normal pregnancies, with reports in the literature ranging from 0%-6%. There is also disagreement as to how much fetal blood contamination affects ACE assay. The test is less affected than AFP assay, but substantial contamination seems capable of producing falsely elevated results.

    Chromosome analysis (cytogenetic karyotyping) on fetal amniotic cells obtained by amniocentesis is the standard way for early prenatal diagnosis of fetal trisomies and other congenital abnormalities. However, standard karyotyping is very time-consuming, requires a certain minimum number of fetal cells that need culturing, and usually takes several days to complete. A new method called fluorescent in situ hybridization (FISH) uses nucleic acid (deoxyribonucleic acid, DNA) probes to detect certain fetal cell chromosomes such as 13, 18, 21, X, and Y, with identification accomplished by a fluorescent dye coupled to the probe molecules. Correlation with traditional cytogenetics has generally been over 95%, with results in 24 hours or less. FISH modifications have made it possible to detect fetal cells in maternal blood and subject them to the same chromosome analysis. One company has a combined-reagent procedure that can be completed in 1 hour. Disadvantages of FISH include inability to detect abnormalities in chromosomes other than the ones specifically targeted by the probes and inability to detect mosaic abnormalities or translocations, thereby missing an estimated 35% of chromosome defects that would have been identified by standard karyotyping methods.

    Preterm labor and placental infection

    It has been estimated that about 7% of deliveries involve mothers who develop preterm labor. It has also been reported that chorioamnionitis is frequently associated with this problem (about 30%; range, 16%-82%). Less than 20% of infected patients are symptomatic. Diagnosis of infection has been attempted by amniotic fluid analysis. Amniotic fluid culture is reported to be positive in about 20% of cases (range, 4%-38%). Mycoplasmas are the most frequent organisms cultured. Amniotic fluid Gram stain is positive in about 20% of patients (range, 12%-64%). Amniotic fluid white blood cell count was reported to be elevated in 57%-64% of cases. However, there was great overlap between patients with or without infection and also between those with proven infection. In three reports, the most sensitive amniotic fluid test for infection was amniotic fluid interleukin-6 (IL-6) assay (81%-100%). However, at present most hospitals would have to obtain IL-6 assay from large reference laboratories.

  • Fetal Maturity Tests

    Tests for monitoring fetal maturity via amniocentesis are also available. Bilirubin levels in erythroblastosis are discussed in chapter 11. Amniotic creatinine assay, amniotic epithelial cell stain with Nile blue sulfate, fat droplet evaluation, osmolality, and the Clemens shake test, alone or in combination, have been tried with varying and not entirely satisfactory results. Most current tests measure one or more components of alveolar surfactant. Surfactant is a substance composed predominantly of phospholipids; it is found in lung alveoli, lowers the surface tension of the alveolar lining, stabilizes the alveoli in expiration, and helps prevent atelectasis. Surfactant deficiency causes neonatal respiratory distress syndrome (RDS), formerly called “hyaline membrane disease.” The major phospholipid components of surfactant are phosphatidylcholine (lecithin, about 80%; range, 73%-88%), phosphatidylglycerol (PG, about 3%; range, 1.8%-4.2%), and sphingomyelin (about 1.6%). The current most widely used tests are the lecithin/sphingomyelin (L/S) ratio, assay of phosphatidylglycerol (PG), the foam stability index (FSI), and TDx fluorescent polarization.

    Lecithin/Sphingomyelin (L/S) ratio. The L/S ratio has been the most widely accepted fetal maturity procedure. Lecithin (phosphatidylcholine), a phospholipid, is the major component of alveolar surfactant. There is a 60% or greater chance of RDS in uncomplicated pregnancies when the fetus is less than 29 weeks old; about 8%-23% at 34 weeks; 0%-2% at 36 weeks; and less than 1% after 37 weeks. In amniotic fluid, the phospholipid known as sphingomyelin normally exceeds lecithin before the 26th week; thereafter, lecithin concentration is slightly predominant until approximately the 34th week, when the lecithin level swiftly rises and the sphingomyelin level slowly decreases so that lecithin levels in the 35th or 36th week become more than twice sphingomyelin levels. After that happens it was originally reported (not entirely correctly), that there was no longer any danger of neonatal RDS. The L/S ratio thus became a test for fetal lung and overall maturity. Certain precautions must be taken. Presence of blood or meconium in the amniotic fluid or contamination by maternal vaginal secretions may cause a false increase in lecithin and sphingomyelin levels, so that “mature” L/S ratios are decreased and “immature” L/S ratios are increased. The amniotic fluid specimen must be cooled immediately, centrifuged to eliminate epithelial cells and other debris, and kept frozen if not tested promptly to prevent destruction of lecithin by certain enzymes in the amniotic fluid.

    Evaluations in unselected amniocentesis patients have revealed that about 55% of neonates with immature L/S ratios using the 2.0 ratio cutoff point do not develop RDS and about 5% (literature range, 0%-17%) of neonates with a mature L/S ratio (ratio >2.0) develop RDS. Some have attempted to eliminate the falsely mature cases by changing the cutoff point to a ratio of 2.5, but this correspondingly increases the number of falsely immature results. In clinically normal pregnancies, only about 3% of neonates with a mature L/S ratio, using proper technique, develop RDS. In complicated pregnancies, especially those with maternal type I insulin-dependent diabetes, hypertension, or premature rupture of the amniotic membrane, about 15% (literature range, 3%-28%) of neonates with mature L/S ratios are reported to develop RDS. In other words, RDS can develop at higher L/S ratios in a relatively small number of infants. The wide range in the literature reflects differences in opinion among investigators as to the effect of diabetes on neonatal L/S ratios. Also, the L/S ratio can produce falsely mature results if contaminated by blood or meconium. It takes experience and careful attention to technical details to obtain consistently accurate L/S results.

    Phosphatidylglycerol (PG). A number of other tests have been developed in search of a procedure that is more accurate in predicting or excluding RDS and that is also technically easy to perform. PG is a relatively minor component (about 10%) of lung surfactant phospholipids. However, PG is almost entirely synthesized by mature lung alveolar cells and therefore is a good indicator of lung maturity. In normal pregnancies PG levels begin to increase after about 30 weeks’ gestation and continue to increase until birth. It normally becomes detectable about the 36th week. In conditions that produce severe fetal stress, such as maternal insulin-dependent diabetes, hypertension, and premature membrane rupture, PG levels may become detectable as early as 30 weeks’ gestation. Most of the limited studies to date indicate that the presence of PG in more than trace amounts strongly suggests that RDS will not develop, whether the pregnancy is normal or complicated. Overall incidence of RDS when PG is present seems to be about 2% (range 0%-10%). It is considered to be a more reliable indicator of fetal lung maturity than the L/S ratio in complicated pregnancy. It may be absent in some patients with clearly normal L/S ratios and occasionally may be present when the L/S ratio is less than 2.0. PG assay is not significantly affected by usual amounts of contamination by blood or meconium.

    PG can be assayed in several ways, including gas chromatography, thin-layer chromatography (TLC), enzymatically, and immunologically. The TLC technique is roughly similar to that of the L/S ratio. Some report the visual presence or absence of PG, with or without some comment as to how much appears to be present (trace or definite). Some report a PG/sphingomyelin (PG/S) ratio. A PG/S ratio of 2.0 or more is considered mature. A commercially available enzymatic PG method (“PG-Numeric”) separates phospholipids from the other components of amniotic fluid (using column chromatography or other means), followed by enzymatic assay of glycerol in the phospholipid fraction. After several years there is still an insufficient number of published evaluations of this technique. Immunological methods are still restricted to a slide agglutination kit called Amniostat FLM-Ultra (improved second-generation test). Current small number of evaluations indicate that Amniostat FLM-Ultra detects about 85%-90% of patients who are positive for PG on TLC. The risk of RDS is about 1%-2% if the test is reactive (positive).

    Foam stability index (FSI). The FSI is a surfactant function test based on the ability of surfactant to lower surface tension sufficiently to permit stabilized foaming when the amniotic fluid is shaken. This depends on the amount and functional capability of surfactant as challenged by certain amounts of the antifoaming agent ethanol. It is thought that the phospholipid dipalmitoyl lecithin is the most important stabilizing agent. The FSI is actually a modification of the Clemens shake test, which used a final amniotic fluid-ethanol mixture of 47.5% ethanol. The FSI consists of a series of seven tubes containing amniotic fluid with increasing percentages of ethanol in 1% increments from 44%-50%. The endpoint is the tube with the highest percentage of ethanol that maintains foam after shaking. An endpoint of the 47% tube predicts about a 4% chance of RDS and an endpoint in the 48% tube predicts less than 1% chance. Because even tiny inaccuracies or fluctuations of ethanol concentration can influence results considerably, and also the tendency of absolute ethanol to adsorb water, some problems were encountered in laboratories making their own reagents. To solve these problems a commercial version of the FSI called Lumidex was introduced featuring sealed tubes containing the 1% increments of ethanol to which aliquots of amniotic fluid are added through the rubber caps that seal the tubes. The FSI (or Lumidex) has been reported to be more reliable than the L/S ratio in predicting fetal lung maturity. At least two reports indicate that the FSI correctly demonstrates fetal lung maturity much more frequently than the L/S ratio in fetuses who are small for their gestational age. Drawbacks of the FSI method in general are interference (false positive) by blood, meconium, vaginal secretions, obstetrical creams, and mineral oil. A major drawback of the current Lumidex kit is a shelf-life of only 3 weeks without refrigeration. Although the shelf life is 3 months with refrigeration, it is necessary to stabilize the tubes at room temperature for at least 3 hours before the test is performed.

    TDx-FLM fluorescent polarization. The TDx is a commercial instrument using fluorescent polarization to assay drug levels and other substances. It has been adapted to assay surfactant quantity indirectly by staining surfactant in amniotic fluid with a fluorescent dye and assaying surfactant (in mg/gm of albumin content) using the molecular viscosity of the fluid as an indicator of surfactant content. The assay in general produces results similar to the L/S ratio and a little better than the FSI. There is some difference in results depending on whether a single cutoff value is used, what that value is, and whether multiple cutoff values are applied depending on the situation. Test technical time is about 30 minutes. Specimens contaminated with meconium, blood, or urine (in vaginal pool material) interfere with the test.

    Lamellar body number density. Surfactant is produced by alveolar type II pneumocytes in the form of a concentrically wrapped small structure about 3 microns in size that on cross-section looks like an onion and is called a lamellar body. It is possible to count the lamellar bodies using some hematology platelet counting machines, with the result calculated in units of particle density per microliter of amniotic fluid. In the very few evaluations published to date, results were comparable to those of the L/S ratio and FSI.

    Amniocentesis laboratory problems. Occasionally, amniotic puncture may enter the maternal bladder instead of the amniotic sac. Some advocate determining glucose and protein levels, which are high in amniotic fluid and low in normal urine. To prevent confusion in diabetics with glucosuria, it has been suggested that urea and potassium levels be measured instead; these are relatively high in urine and low in amniotic fluid. Another potential danger area is the use of spectrophotometric measurement of amniotic fluid pigment as an estimate of amniotic fluid bilirubin content. Before 25 weeks’ gestation, normal pigment levels may be greater than those usually associated with abnormality.

  • Fetal or Placental Function

    Urine estriol or total estrogens. Estriol is an estrogenic compound produced by the placenta from precursors derived from fetal adrenal cortex and fetal liver. Newly synthesized estriol is unconjugated; therefore, unconjugated estriol represents a product of the entire fetoplacental unit. The unconjugated estriol reaches maternal serum (where it has a half-life of about 20 minutes) and is taken to the maternal liver, where about 90% is conjugated with a glucuronide molecule. The conjugated form of estriol is excreted in maternal urine. A lesser amount of conjugated estriol is produced by the maternal liver from nonestriol estrogens synthesized by the placenta from maternal adrenal precursors. Serum estriol can be measured either as total estriol or as unconjugated estriol. It usually is measured as unconjugated estriol to exclude maternal contribution to the conjugated fraction. Urine estriol can be measured as total estriol or as total estrogens, since estriol normally constitutes 90%-95% of urine total estrogens.

    Historically, urine total estrogen was the first test used, since total estrogen can be assayed by standard chemical techniques. However, urine glucose falsely increases results (which is a problem in diabetics, who form a large segment of the obstetrical high-risk group), and certain other substances such as urobilinogen also may interfere. In addition, urine total estrogen results are a little more variable than urine estriol patterns. Eventually, other biochemical procedures that were more specific for urine estriol (in some cases, however, the “estriol” being measured using biochemical methods is actually total estrogen) were devised. These procedures also have certain drawbacks, some of which are shared by the total estrogen methods. Both urine total estrogens and urine estriol have a significant degree of between-day variation in the same patient, which averages about 10%-15% but which can be as high as 50%. Both are dependent on renal excretion, and, therefore, on maternal renal function. Both have a maternal component as well as the fetal component. Urine total estrogen necessitates a 24-hour collection. The standard method for urine estriol also is a 24-hour specimen. There is substantial difficulty collecting accurate 24-hour specimens, especially in outpatients. Also, there is a 1- to 2-day time lag before results are available. Some have proposed a single voided specimen based on the estriol/creatinine ratio. However, there is controversy whether the single-voided specimen method (reported in terms of either estriol per gram creatinine or estriol/creatinine ratio) provides results as clinically accurate as the 24-hour specimen.

    Estriol can be detected by immunoassay as early as the ninth week of gestation. Thereafter, estriol values slowly but steadily increase until the last trimester, when there is a more pronounced increase. Clinical use of estriol measurement is based on the fact that severe acute abnormality of the fetoplacental unit (i.e., a dead or dying placenta or fetus) is manifested either by failure of the estriol level to continue rising or by a sudden marked and sustained decrease in the estriol level. Urine specimens are usually obtained weekly in the earlier part of pregnancy, twice weekly in the last trimester, and daily for several days if a problem develops.

    In general, urine estriol or estrogen excretion correlates reasonably well with fetal health. However, there are important exceptions. Only severe fetal or placental distress produces urine estrogen decrease of sufficient magnitude, sufficient duration, and sufficiently often to be reliable (i.e., mild disorder may not be detected). There is sufficient daily variation in excretion so that only a very substantial and sustained decrease in excretion, such as 40%-50% of the mean value of several previous results, is considered reliable. Some consider an estriol value less than 4 mg/24 hours (after 32 weeks’ gestation) strongly suggestive of fetal distress and a value more than 12 mg/24 hours as indicative of fetal well-being. Maternal hypertension, preeclampsia, severe anemia, and impaired renal function can decrease urine estrogen or estriol excretion considerably. Decrease may also occur to variable degree in variable numbers of fetuses with severe congenital anomalies. Certain drugs such as ampicillin and cortisol may affect urine estriol or estrogen values by effects on production, and other substances such as mandelamine or glucose can alter results from biochemical interference with some test methods. Some investigators have reported a decrease shortly before delivery in a substantial minority of normal patients. Maternal Rh-immune disease may produce a false increase in urine estriol levels. Continued bed rest has been reported to increase estriol excretion values an average of 20% over levels from ambulatory persons, with this increase occurring in about 90% of patients in the third trimester.

    The literature contains widely differing opinions regarding clinical usefulness of estrogen excretion assay in pregnancy. In general, investigators have found that urinary estrogen or estriol levels are decreased in about 60%-70% of cases in which fetal distress occurs (literature range, 33%-80%). The more severe the fetal or placental disorder, the more likely that urine estrogen or estriol levels will be low. The percentage of falsely low values is also said to be substantial, but numerical data are not as readily available.

    Serum unconjugated estriol. Plasma or serum unconjugated estriol, measured by immunoassay, has been used as a replacement for urine hormone excretion. Advantages include ease of specimen collection (avoidance of 24-hour urine collection problems), increased specificity for fetoplacental dysfunction (no maternal hormone contribution), no 24-hour wait for a specimen, closer observation of fetoplacental health (due to the short unconjugated estriol half-life), little technical interference by substances such as glucose, and less dependence on maternal kidney function. Drawbacks include the majority of those drawbacks previously described for urine hormone excretion (effect of bed rest, hypertension, and other conditions, and medications affecting estrogen production). Also similar to urine excretion, there is substantial between-day variation, averaging about 15% (reported maximum variation up to 49%). However, there are also considerable within-day fluctuations, which average about 15% (with maximum variation reported as high as 51%). Thus, a single value is even more difficult to interpret than a urine value. Some believe that 24-hour urine measurements may thus have some advantage, since within-day fluctuations are averaged out. Also similar to urine values, the current trend of interpretation is to require a sustained decrease of 40%-50% from the average of several previous serum values to be considered a significant abnormality. The serum specimens should preferably be drawn at the same time of day in the same patient position and assayed by the same laboratory. Although frequency of sampling is not uniform among investigators, many obtain one or two specimens per week during the earlier part of pregnancy and one per day if there is clinical suggestion of abnormality or one serum value becomes significantly decreased. Although there is some disagreement, the majority of investigators indicate that serum unconjugated estriol has a little better correlation with clinical results than does urine hormone excretion.

    Because of the problems associated with collection of urine or serum estriol specimens and interpretation of the values, as well as the disturbing number of false positive and false negative test results, many clinicians depend more on other procedures (e.g., the nonstress test, which correlates the rate of fetal heartbeat to fetal movement) than on estrogen values to monitor fetal well-being.

    Placental lactogen. HPL is a hormone produced only by the placenta, with metabolic activity similar in some degree to that of prolactin and GH. Values correlate roughly with the weight of the placenta and rise steadily in maternal serum during the first and second trimesters before entering a relative plateau in the third. Serum levels of hPL are higher than those attained by any other peptide hormone. Serum half-life is about 30 minutes. Although hPL cross-reacts with GH in most radioimmunoassay (RIA) systems, the high level of hPL relative to GH at the time of pregnancy when hPL levels are measured prevents clinical problems with GH interference. Serum hPL has been evaluated by many investigators as a test of placental function in the third trimester. Its short half-life is thought to make it a more sensitive indicator of placental failure than measurements of other hormones, especially urine measurements. Since hPL levels normally can fluctuate somewhat, serial measurements are more accurate than a single determination. Estriol, which reflects combined fetal and placental function, still seems to be used more than hPL.

  • Pregnancy Tests

    Most pregnancy tests are based on the fact that the placenta secretes human chorionic gonadotropin (hCG), a hormone that has a luteinizing action on ovarian follicles and probably has other functions that are not completely known. Serum hCG levels of about 25 milli-international units (mIU)/ ml (IU/L) are reached about 8-10 days after conception. The hCG levels double approximately every 2 days (various investigators have reported doubling times ranging from 1-3 days) during approximately the first 6 weeks of gestation. Levels of about 500 mIU/ml are encountered about 14-18 days after conception (28-32 days after the beginning of the last menstrual period). Serum levels are generally higher than urine levels for about the first 2 weeks after conception and about the same as urine levels during the third week. Thereafter, urine levels are higher than serum. The serum (and urine) hCG levels peak about 55-70 days (8-10 weeks) after conception (literature range, 40-77 days). Peak serum values are about 30,000 mIU/ml (range, 20,000-57,000 mIU/ml). Serum and urine levels then decline rather rapidly during the last part of the first trimester, with serum levels eventually stabilizing at about 10,000 mIU/ml. These levels are maintained for the remainder of pregnancy, although some investigators describe a brief rise and fall in the third trimester. Urine levels generally parallel serum levels, but the actual quantity of urine hCG obtained in terms of milliinternational units per milliliter is considerably dependent on technical aspects of the kit method being used (discussed later).

    The hCG molecule is composed of two subunits, alpha and beta. The alpha subunit is also a part of the pituitary hormones LH, FSH, and TSH. The beta subunit, however, is different for each hormone. The hCG molecule in serum becomes partially degraded or metabolized to beta subunits and other fragments that are excreted in urine.

    Biologic tests. The first practical biologic test for pregnancy was the Ascheim-Zondek test, published in 1928. Urine was injected into immature female mice, and a positive result was indicated by corpus luteum development in the ovaries. This took 4-5 days to perform. The next major advance took place in the late 1950s when frog tests were introduced. These took about 2 hours to complete. The result was almost always positive by the 40th day after the last menses, although it could become positive earlier.

    Immunologic tests. In the 1960s it was learned that antibodies to hCG could be produced by injecting the hCG molecule into animals. This was the basis for developing immunologic pregnancy tests using antigen-antibody reactions. In the late 1960s and during the 1970s, both latex agglutination slide tests and hemagglutination tube tests became available. The slide tests took about 2 minutes to perform and had a sensitivity of 1,500-3,000 mIU/ml, depending on the manufacturer. The tube tests required 2 hours to complete and had a sensitivity of 600-1,500 mIU/ml. The antibody preparations used at that time were polyclonal antibodies developed against the intact hCG molecule, and they cross-reacted with LH and TSH. This did not permit tests to be sensitive enough to detect small amounts of hCG, because urine LH could produce false positive results.

    Beta subunit antibody tests. In the late 1970s, methods were found to develop antibodies against the beta subunit of hCG rather than against the whole molecule. Antibody specific against the beta subunit could greatly reduce or even eliminate the cross-reaction of hCG with LH. However, the degree of current beta subunit antibody specificity varies with different commercial companies. By 1980, sensitivity of the slide tests using beta hCG antibody had reached 500-1,500 mIU/ml, and sensitivity of the beta hCG tube tests was approximately 200 mIU/ml. Both the slide and the tube tests required a urine specimen. In the 1980s, standard immunoassay methods were developed for beta hCG in serum that provide a sensitivity of 3-50 mIU/ml. These methods detect pregnancy 1-2 weeks after conception. The great majority of current tests use monoclonal antibodies, either alone or with a polyclonal antibody that captures the hCG molecule and a monoclonal antibody that identifies it. Several manufacturers developed abbreviated serum pregnancy immunoassays that compared patient serum with a single standard containing a known amount of beta hCG (usually in the range of 25 mIU/ml). A result greater than the standard means that beta hCG is present in a quantity greater than the standard value, which in usual circumstances indicates pregnancy. Current serum immunoassay procedures take between 5 minutes and 2 hours to perform (depending on the manufacturer). The abbreviated method is much less expensive and is usually quicker. Several urine tests are available that detect 50 mIU/ml of hCG.

    Technical problems with human chorionic gonadotropin. Some (not all) of the kits that detect less than 25 mIU/ml of hCG may have problems with false-positive results of several different etiologies. First, of course, there may be incorrect performance of the test or patient specimen mishandling. The antibodies used in the different manufacturer’s tests have different specificities. Serum hCG molecules may exist in different forms in some patients: whole (“intact”) molecule, free beta subunit, free alpha subunit, or other degraded hCG fragments. Considerable quantities of serum free beta or alpha subunits are more often seen with tumors. Different antibodies may detect different amounts of hCG material depending on whether the antibody detects only the whole molecule, the beta subunit on the whole molecule, or the free beta subunit only (including in urine a partially degraded free beta subunit known as the “core beta fragment”). Most anti-beta antibodies actually detect both whole molecule (because of the structural beta subunit), free beta subunit, and core beta fragments. Therefore, the amount of hCG (in mIU/ml) detected in urine depends on several factors: (1) whether a specific whole-molecule or a beta-hCG method is used. The specific whole-molecule method reports about the same quantity of intact hCG in serum or urine, whereas the beta-specific assay would report higher amounts of hCG in urine than in serum since it detects intact hCG plus the beta subunits and fragments that are present in greater quantities in urine than serum; (2) degree of urine concentration or dilution; (3) stage of pregnancy, since more beta fragments appear in urine after the first few weeks; (4) how the particular kit is standardized (discussed later). Some beta-hCG antibodies have a certain degree of cross-reaction with LH, although theoretically a beta-specific antibody should not do so. The serum of occasional patients contains heterophil-type antibodies capable of cross-reacting with monoclonal test antibodies (HAMA) that were produced in mouse tissue cells and could produce a false positive result. This most often happens with double antibody “sandwich” test methods. Some kits are affected in a similar way by renal failure.

    Another confusing aspect of pregnancy testing relates to standardization of the tests by the manufacturers (that is, adjusting the test method to produce the same result as a standard, which is a known quantity of the material being assayed). In hCG testing, the manufacturers use a standard from the World Health Organization (WHO). The confusion arises because the earliest WHO standard used for this purpose (Second International Standard; second IS) was composed of a mixture of whole-molecule hCG, free beta subunits, and other hCG fragments. When the supply of second IS was exhausted, the WHO developed the first (and then the third) International Reference Preparation (IRP), which is mostly whole-molecule hCG without free beta subunit. However, hCG kits standardized with the original second IS give results about half as high as current kits standardized against the first or third IRP. Also, certain current kits specific for whole-molecule hCG would not detect some of the hCG fragments in the original second IS. This difference in antibody behavior may at least partially explain discrepant reports in the literature of equal quantities of hCG in pregnancy serum and urine and other reports of urine values as high as 10 times serum values. After the first few weeks of pregnancy, maternal serum contains primarily intact hCG; maternal urine contains some intact hCG but much larger quantities of free beta subunits and core beta fragments.

    Finally, it has been reported in several studies that occasionally normal nonpregnant women may have low-level circulating levels of an hCG-like substance, usually less than 25 mIU/ml. This was reported in about 2.5% (range, 0%-14%) of patients in these studies, although most evaluations of hCG test kits have not reported false positive results. When one kit was reactive, sometimes one or more different kits would also be reactive, but usually some kits do not react with these substances. At present, in most laboratories there is no satisfactory way to know immediately whether a positive result is due to pregnancy, is due to hCG-producing tumor, or is false positive, especially when the test is a yes-no method. Although there are ways to investigate possible discrepancies, it usually takes considerable time and retesting to solve the problem or it may necessitate consultation with a reference laboratory.

    Other uses for hCG assay. Pregnancy tests are useful in certain situations other than early diagnosis of normal pregnancy. These conditions include ectopic pregnancy, spontaneous abortion (which occurs in about 15% of all pregnancies; literature range, 12%-31%), and hCG-producing neoplasms. Ectopic pregnancy and neoplasms will be discussed in detail later. When the differential diagnosis includes normal intrauterine pregnancy, ectopic pregnancy, and threatened, incomplete, or complete abortion, the pattern obtained from serum quantitative beta-hCG assays performed every other day may be helpful. During the first 4 weeks of pregnancy (beginning at conception), there is roughly a doubling of hCG every 2 days (range, 1-3 days). As noted earlier, serum beta hCG by immunoassay first detects embryonic placental hCG in titers of 2-25 IU/L between 1 and 2 weeks after conception. Ectopic pregnancy and abortions may demonstrate an increase in their hCG levels at the same rate as in normal pregnancy up to a certain point. In the case of ectopic pregnancy, that point is usually less than 4 weeks (possibly as long as 6 weeks) after conception, since the ectopic location usually limits placental growth or rupture occurs. The typical pattern of ectopic pregnancy is a leveling off (plateau) at a certain time. The usual pattern of abortion is either a decrease in beta-hCG levels as abortion takes place or a considerable slowing in the rate of increase. However, these are only rules of thumb. About 15% of normal intrauterine pregnancies display less increase (decreased rate of increase) than expected, and thus could be mistaken for beginning abortion by this criterion alone. Also, ectopic pregnancy values may sometimes decline rather than plateau if the fetus dies.

    Ectopic pregnancy

    Ectopic pregnancy is a common gynecologic problem, either by itself or in differential diagnosis. Symptoms include abdominal pain of various types in about 97% of patients (literature range, 91%-100%), abnormal uterine bleeding in about 75% (54%-80%), delayed menses in about 75% (68%-84%), adnexal tenderness on palpation in about 90%-95%, unilateral adnexal mass in about 50% (30%-76%), and fever (usually lowgrade) in about 5% (3%-9%). Hypovolemic shock is reported as the presenting symptom in about 14%. It is obvious that these signs and symptoms can suggest a great number of conditions. In one study, 31% of patients with ectopic pregnancy in the differential diagnosis had a strongly suggestive triad of abdominal pain, uterine bleeding, and an adnexal mass. Only 14% of these patients were found to have ectopic pregnancy. Some conditions that frequently mimic ectopic pregnancy are pelvic inflammatory disease; threatened, incomplete, or complete abortion; corpus luteum rupture; dysfunctional uterine bleeding; and bleeding ovarian cyst. Among routine laboratory tests, a hemoglobin value less than 10 gm/100 ml is reported in about 40% of ectopic pregnancy cases (28%-55%) and leukocytosis in about 50%. Among other diagnostic procedures, culdocentesis for fresh blood is reported to produce about 10% false negative results (5%-18%). Pregnancy test results vary according to the sensitivity of the test. Urine or serum pregnancy tests with a sensitivity of 500-1,000 mIU/ml result in about 25% false negative results (8%-60%). Tests with a sensitivity of 50 mIU/ml yield about 5%-10% false negative results (0%-13%). Serum tests with a sensitivity of 25 IU/L or better have a false negative rate of about 1%-2% (range, 0%-3%). A positive pregnancy test result is not a diagnosis of ectopic pregnancy but signifies only that the patient has increased levels of hCG, for which there could be several possible causes. Also, some manufacturers’ kits are subject to a certain number of false positive results. Interpretation of a negative test result depends on the sensitivity of the test. If the test is a serum hCG immunoassay with a sensitivity of 25 mIU/ml (IU/L) or better, a negative test result is about 98%-99% accurate in excluding pregnancy. However, there are rare cases in which the specimen might be obtained 2-4 days before the patient hCG reaches detectable levels or there could be a technical laboratory error. A repeat test 48 hours later helps to exclude these possibilities.

    As noted previously, failure to double hCG values in 24 hours at gestational age 4-8 weeks occurs in about 66% of ectopic pregnancies, about 85% of spontaneous abortion cases, and about 15% of normal pregnancies. Such an abnormally slow hCG increase rate would warrant closer followup or possibly other diagnostic tests, such as a quantitative serum hCG assay if the 48-hour increase is considerably low. A substantially low serum hCG level for gestational age suggests abnormal pregnancy. Another use for quantitative hCG assay in appropriate cases is to see if the “discriminatory zone” of Kadar has been reached. Originally, this was the range of 6,000-6,500 mIU/ml (IU/L, IRP standard) above which standard transabdominal ultrasound (TAUS) can visualize a normal pregnancy gestational sac in the uterus in about 94% of cases (although TAUS could detect an intrauterine gestational sac below 6,000 mIU/ml in some cases, failure to do so gives no useful information). With more sensitive ultrasound equipment and use of a vaginal transducer, it has been reported that the discriminatory zone upper limit can be reduced to the area of 1,000-1,500 mIU/ml (IU/L), but the exact value must be established by each institution using its particular pregnancy test and ultrasound equipment. Transvaginal ultrasound is more sensitive than TAUS in detecting an adnexal mass or free cul-de-sac fluid that would suggest ectopic pregnancy.

    Neoplasms producing human chorionic gonadotropin

    Neoplasms arising from chorionic villi, the fetal part of the placenta, are known as gestational trophoblastic neoplasms and include hydatidiform mole (the counterpart in tumor classification of benign adenoma) and choriocarcinoma (chorioepithelioma, the equivalent of carcinoma). Hydatidiform mole also has a subdivision, chorioadenoma destruens, in which the neoplasm invades the placenta but there is no other evidence of malignancy. The major importance of hydatidiform mole is a very high (і 10%) incidence of progression to choriocarcinoma.

    Several hormone assays have been proposed as aids in diagnosis. By far the most important is hCG, which is produced by the trophoblast cell component of fetal placental tissue. Current pregnancy tests using monoclonal antibodies to beta subunit of hCG or to the whole molecule can detect levels of 25 mIU/ml (IU/L), sometimes less, without interference by LH, permitting detection of nearly all gestational tumors (except a very few that predominately secrete the free beta fragment of hCG, which would necessitate an assay that would detect this hCG metabolite). Since normal placental tissue secretes hCG, the problem then is to differentiate normal pregnancy from neoplasm. Suspicion is raised by clinical signs and also by finding hCG levels that are increased more than expected by the duration of pregnancy or that persist after removal of the placenta. Twin or other multiple pregnancies can also produce hCG levels above expected values. Although serum levels of hCG greater than 50,000 mIU/ml (or urine levels > 300,000 mIU/ml) are typically associated with gestational neoplasms, especially if these levels persist, a considerable number of patients with gestational tumors have hCG values less than this level. About 25% of patients in one report had values less than 1,000 mIU/ml. In normal pregnancies, serum hCG levels become nondetectable by about 14 days (range, 3-30 days) after delivery. In one study of elective abortions, it took 23-52 days for hCG levels to become nondetectable. After removal of a hydatidiform mole, hCG levels should become nondetectable in about 2-3 months (range, 21-278 days). Once neoplasm is diagnosed and treated, hCG measurement is a guideline for success of therapy and follow-up of the patient for possible recurrence.

    Other hormones useful in possible gestational neoplasms. Fetal and placental tissue produces other hormones that may be useful. Progesterone (or its metabolite pregnanediol) and estradiol are secreted by the placenta in slowly increasing quantity throughout most of pregnancy. It has been reported that during the first 20 weeks of gestation, hydatidiform moles are associated with serum estradiol-17b values that are increased from values expected in normal pregnancy, with good separation of normal pregnancy from molar pregnancy. Serum progesterone levels were increased in about 75% of nonaborted moles up to the 20th week. Urinary pregnanediol levels, on the other hand, are frequently decreased. Finding increased serum progesterone and estradiol-17b levels during the time that peak hCG values are expected (between the 50th and 80th days after the last menstrual period), accompanied by a decreased urine pregnanediol level, would suggest a hydatidiform mole or possibly a choriocarcinoma. Serum human placental lactogen (hPL), or somatomammotropin, is another placental hormone whose level rises during the first and second trimesters and then reaches a plateau during the last 2-3 months. The association of decreased levels of hPL in the first and second trimesters with increased hCG levels suggests neoplasm. There is, however, a small degree of overlap of hPL level in patients with mole and the normal range for pregnancy. One report suggests a possible inverse ratio between hPL values and degree of malignancy (the greater the degree of malignancy, the less serum hPL produced).

    Production of hCG has been reported to occur in nearly two thirds of testicular embryonal cell carcinomas and in about one third of testicular seminomas. Instances of hCG secretion by adenocarcinomas from other organs and, rarely, from certain other tumors have been reported.