Recent studies examined the value of aPL in predicting adverse fetal and maternal outcome in healthy pregnant women [20-24]. The correlation of aPL with low birth weight, intrauterine fetal death, and adverse complications of pregnancy was identified by some authors [21, 24-26] but not by others [22]. These studies concentrated on the predictive value of aPL as measured by anticardiolipin antibody (IgG, IgM) or lupus anticoagulant (activated partial thromboplastin time). Studies have not examined the individual and overlapping value of multiple measures of aPL in predicting pregnancy outcome.

We began this study primarily to investigate the association between the presence of aPL at the first prenatal visit and adverse pregnancy outcome in healthy pregnant women. We also examined the confounding effects of covariables and other potential pregnancy risk factors. To understand the natural history and temporal variation of these antibodies, we also measured aPL at delivery in a subgroup of our patients. We focused on clinically low-risk pregnant women in whom aPL would not be routinely requested. Thus, we chose women who had never been pregnant before or nulligravida women who had fewer than two spontaneous abortions. By adopting this approach, we eliminated the possible confounding effect of previous complications affecting pregnancy outcome and avoided the ethical dilemma of potential therapeutic intervention in the event of an aPL-positive test result.

Methods

Top Methods Results Discussion Author & Article Info References

Study Design

We conducted the study at the University of Colorado Health Sciences Center [UCHSC] after obtaining approval from the institutional review board. From August 1990 to July 1991, we invited all consecutive nulliparous patients seen in the Obstetrics Clinic before 25 weeks gestation to participate. Exclusion criteria were 1) a history of more than two spontaneous abortions; 2) a history of systemic lupus erythematosus; 3) current treatment with corticosteroids or immunosuppressive agents; and 4) a history of thrombosis. We interviewed 451 women who met the inclusion criteria. Forty of them chose not to participate, and 3 were subsequently removed from the study (the fetus of 1 woman died in utero after maternal trauma and 2 women had therapeutic abortions). Four hundred and eight women were enrolled for prospective study. We obtained informed consent from all women at the first visit. We gathered data on age, race, marital status, and medical and obstetric history from the case records. Gestational age was determined by last menstrual period, fundal height, or ultrasound examination (when menstrual dates and fundal height did not concur). Thereafter, participants were followed by their attending physicians throughout pregnancy and at delivery. All pregnancies were completed by April 1992.

We reviewed case records when each pregnancy was completed. We collected data on the participants' use of cigarettes before and after conception and on the following defined pregnancy complications [27]: fetal loss (subdivided into before and after 20 weeks gestation), neonatal loss, pregnancy-induced hypertension, preeclamptic toxemia, abruptio placentae, placenta previa, gestational diabetes, premature rupture of the membranes, preterm labor, and preterm delivery and low birth weight. We noted Apgar scores at 5 minutes, birth weight, and gestational age at delivery. We collected follow-up data by mail and telephone for all patients who subsequently delivered outside UCHSC. The latter group included patients who moved to another hospital, moved out of state, or were lost to medical review after an early fetal loss at home. Of the 408 women originally enrolled, 389 had blood drawn at the first visit and completed clinical follow-up for maternal and fetal complications. This resulted in a 4.7% loss to clinical follow-up. Of the 389 patients who had blood drawn initially, 239 had samples drawn at delivery. The loss of samples at delivery was due to delivery at other sites, failure to follow study protocol (collect blood), or early fetal loss occurring in a nonmedical setting.

Blood Sample Processing

Blood samples were obtained at the first prenatal visit for activated partial thromboplastin time, dilute Russell viper venom time, rapid plasma reagin, and anticardiolipin antibody (IgG, IgM, and IgA) and at delivery for dilute Russell viper venom time and the three isotypes of anticardiolipin antibody. To ensure that all investigators reviewing charts for pregnancy outcome were blinded to aPL results, dilute Russell viper venom time and anticardiolipin antibody measurements were not done until after completion of pregnancy. Blood samples were drawn for serum and plasma by standard techniques. Platelet-poor plasma (<5000 platelets/µL) was obtained by centrifugation at 3300g to remove the cell mass and platelets. An activated partial thromboplastin time test was done immediately on an aliquot of plasma from the first prenatal visit. The remaining plasma was frozen at –70°C for subsequent testing for dilute Russell viper venom time. Serum was separated from the cell mass and frozen at –70°C to determine the anticardiolipin antibody isotypes by enzyme-linked immunosorbent assay (ELISA).

Evaluation of Anticardiolipin Antibodies

Anticardiolipin antibodies (IgG, IgM, and IgA) were measured in duplicate using the commercial anticardiolipin antibody ELISA kit (REAADS Medical Products, Westminster, Colorado). Briefly, serum samples, control sera, and calibrator sera were diluted 1:50 in sample diluent (0.01 mol/L phosphate-buffered saline, pH 7.4, containing 10% bovine serum). One hundred µL of diluted serum was incubated in duplicate wells of 96-well plates coated with beef heart cardiolipin (diphosphatidyl glycerol) for 15 minutes at room temperature. The serum was removed and the wells were washed four times with phosphate-buffered saline. One hundred µL of a prediluted horseradish peroxidase-conjugated goat antihuman IgG or IgM or rabbit anti-IgA was added to each well and incubated for 15 minutes at room temperature. After washing, 100 µL of a 1:1 substrate mixture of tetramethylbenzidine and H₂O₂ was added to the microwells and incubated for 10 minutes at room temperature. Color development was stopped by addition of 100 µL of 2.5 N H_2SO4 and the absorbance at 450 nm was measured using an Emax microtiter plate reader (Molecular Devices, Menlo Park, California). The values of anticardiolipin antibody activity for each sample were calculated from calibrator sera according to the manufacturer's instructions. This assay was standardized relative to the international reference preparations (original 1984 set) obtained from the Anti-Phospholipid Standardization Laboratory, University of Louisville, Louisville, Kentucky. The normal cutoff value of the assay (23 units IgG, 11 units IgM) was defined as the mean IgG or IgM units plus two standard deviations of a healthy patient group. Immunoglobulin A anticardiolipin antibody activity was assessed using REAADS IgA anticardiolipin antibody test kit. A positive level (22 units) was defined as the mean plus three standard deviations of a healthy patient group.

Rapid plasma reagin was measured by standard procedure. For women who tested positive, fluorescent Treponema antibody confirmatory tests were done at the Colorado State Department of Health Laboratory. If these test results were negative, the rapid plasma reagin was labeled false-positive.

Evaluation of Activated Partial Thromboplastin Time

Platelet-poor plasma was incubated for 3 minutes at 37 °C with activated partial thromboplastin time reagent (Auto-aPT; Organon Teknika, Durham, North Carolina) and the time to clot formation was measured by standard procedure (coagumate/2000, Organon Teknika). A value that was 2 or more standard deviations above the mean for healthy controls was considered abnormal.

Evaluation of Dilute Russell Viper Venom Time

Platelet-poor plasma (100 µL) was added to a mixture of dilute phospholipid (1/400 dilution Rabbit Brain Cephalin; Sigma Chemicals, St. Louis, Missouri) and 200 µL Russell's Viper Venom (1/40 000 dilution of 1 mg crude venom; Sigma Chemicals). After 1 minute, 100 µL calcium ions (0.25 mmol/L) was added and the time for clot formation was determined. The established range, based on frozen samples from 50 healthy, nonsmoking, nonpregnant persons of both sexes taking no medications was 26 to 32 seconds (mean ±2 SD).

Statistical Analysis

An aPL profile was created from the six aPL tests done at the first prenatal visit. This profile was called abnormal for a woman if she had one or more abnormal results on any test in the profile. Classification based on this profile was the primary risk variable examined. The primary outcome variable was fetal death. Because the number of abnormal profiles was small (95 of 389), confounding variables (age, sex, race, and so on) were first examined singly using relative risk, chi-square, and Fisher exact tests to determine their effect on either the risk or outcome variable. Any variables whose chi-square probability values were less than 0.1 for association with either the risk or outcome variable were entered into a logistic regression analysis to adjust for potential confounding variables. Finally, the primary risk variable was broken down into its individual test components and the relation of these to the outcome variable was examined using relative risk, chi-square, and Fisher exact tests. A similar profile, but without the activated partial thromboplastin time and rapid plasma reagin, was created at delivery. This delivery profile was compared with fetal loss and with pregnancy complications in a manner similar to the primary risk variable described above.

Those women who had more than one aPL-positive test result at either the prenatal or the delivery examination were analyzed to determine if their probability of experiencing fetal loss or a complication of pregnancy was higher than for those women who had only one positive test result. Those women who were positive at both the prenatal visit and at delivery were examined in a similar manner. Finally, the joint distribution of test scores at the prenatal visit and delivery was examined for an indication of systematic shift. All analyses were done using the SAS statistical package (Cary, North Carolina).

Results

Top Methods Results Discussion Author & Article Info References

Patient Demographics

The mean maternal age at study entry was 21.55 years and the mean gestational age when the first blood was draw was 13.15 weeks. Most of our study population were single mothers (75.8%). The racial distribution was 73.3% white, 15.4% black, and 6.7% Hispanic; other races accounted for 4.6% (Table 1).

Frequency of Different Measures of Antiphospholipid Antibodies

The frequency of positive aPL results at the first prenatal visit and at delivery is shown in Table 2. The rates of aPL positivity at delivery are based only on patients who delivered at UCHSC and had blood drawn at the time of delivery. Some patients who were positive at initial presentation were negative at delivery and vice versa. Consequently, the prevalence rates at the initial visit and at delivery do not necessarily represent the same women. Results of McNemar paired tests done on the 239 women for whom an aPL profile was obtained at both the prenatal and the delivery visit were not significant (P = 0.14), indicating no significant shift in the aPL status of specific women between the prenatal visit and delivery. However, 33 women changed from a positive aPL profile at the first visit to a negative profile at delivery, and 34 women changed from a negative profile at the first prenatal visit to a positive profile at delivery. These data suggest that considerable variation may exist in serial aPL results during pregnancy.

Relation of Antiphospholipid Antibody to Fetal Loss

A positive result on any component of the aPL profile at the first prenatal visit was significantly associated with fetal loss (Table 3). Of the aPL profile-positive women, 15.8% had fetal loss, compared with 6.5% of the aPL profile-negative women (relative risk, 2.44; 95% CI, 1.29 to 4.62; P = 0.011). The relative risk of an adverse fetal outcome was increased for all measures of aPL except IgM anticardiolipin antibody and IgA anticardiolipin antibody. However, the only single aPL test that was significantly associated with fetal loss was IgG anticardiolipin antibody (relative risk, 3.54; CI, 1.56 to 8.07; P = 0.014). A positive result on the aPL profile or any individual measure of aPL at delivery was not significantly associated with fetal loss (data not shown). Thirteen patients had a history of more than one spontaneous abortion. A history of abortion was not significantly associated with fetal loss or the presence of aPL.

Maternal age, race, cigarette use during pregnancy, and gestational age at first prenatal visit were associated with either the primary risk (aPL profile) or primary outcome (fetal loss) variables. All of these confounding variables were therefore entered into a single logistic regression with the primary risk variable. Even when adjusted for these confounding variables, the adjusted relative risk of a positive aPL profile for fetal loss remained significant (adjusted relative risk, 1.57; CI, 1.08 to 2.29). Similarly, the adjusted relative risk of IgG anticardiolipin antibody was 1.92 (CI, 1.07 to 3.44). Having more than one positive aPL test result at either visit, having a positive test result at both the prenatal visit and at delivery, or both did not significantly increase the risk for fetal loss.

Six fetuses died in the anticardiolipin antibody-positive groups. Five were positive for IgG anticardiolipin antibody and 1 was positive for IgM anticardiolipin antibody. These fetal losses were all in women whose anticardiolipin antibody levels were in the low-positive range (23 to 40 IgG antiphospholipid units; 11 to 20 IgM antiphospholipid units). Four of the fetal losses associated with IgG anticardiolipin antibody occurred before 20 weeks gestation and the other was an intrauterine death at 28 weeks gestation. The only fetal loss in the IgM anticardiolipin antibody-positive women occurred at 23 weeks gestation. Only one patient tested positive for IgA anticardiolipin antibody and her pregnancy was uncomplicated. Six fetuses died before 20 weeks gestation in women whose dilute Russell viper venom time values were in the low-positive range (2 to 2.5 SD above the mean). Three women had a prolonged activated partial thromboplastin time (range, 34 to 39 seconds) and experienced fetal loss before 20 weeks gestation.

Relation of Antiphospholipid Antibody to Maternal Complications of Pregnancy

The number of pregnancy-associated complications was small (Table 4). We found no significant association between aPL at the first prenatal visit or at delivery and any complications of pregnancy.

Birth Weights and Apgar Scores

Birth weights at delivery of live-born children were examined as a function of aPL positivity at first prenatal visit and at delivery. The mean birth weight of neonates born to mothers who were aPL positive at the first prenatal visit was slightly higher than the mean birth weight of neonates born to mothers who were aPL negative (3343 g compared with 3238 g; P = 0.19). A higher mean birth weight was noted in the aPL-positive group for each measure of the aPL profile at first visit and at delivery. This trend for increased birth weight in aPL-positive patients was statistically significant in patients with a positive activated partial thromboplastin time test result at the first prenatal visit (P = 0.01) and in patients positive for IgG anticardiolipin antibody at delivery (P = 0.04). Neonates born to mothers who smoked during pregnancy had a lower mean birth weight than those born to nonsmokers (P = 0.04). Apgar scores were considered abnormal if they were less than 7 at 5 minutes. There was no significant association between an abnormal aPL profile or any abnormal aPL test result at the first prenatal visit or at delivery and low Apgar scores. Mean Apgar scores were not significantly different among neonates born to aPL-positive (mean, 8.84 ±0.43) and aPL-negative mothers (mean, 8.47 ±1.6).

Discussion

Top Methods Results Discussion Author & Article Info References

We conducted a prospective study to determine the prevalence and risk of elevated aPL levels in healthy, nulliparous pregnant women. Our data show that the prevalence of individual measures of aPL at the first prenatal visit in healthy pregnant women ranges from 0.26% (IgA anticardiolipin antibody) to 13.68% (dilute Russell viper venom time). These figures are consistent with previous studies that estimated the prevalence of anticardiolipin antibody (IgG and IgM) positivity in healthy women to be 2% to 7% [21, 22]. The most striking finding in our study was the significant association between aPL positivity at the first prenatal visit and fetal loss. All measures of aPL except IgM anticardiolipin antibody and IgA anticardiolipin antibody showed an increased relative risk for fetal loss. This increased risk was statistically significant for an abnormal aPL profile and for IgG anticardiolipin antibody. A false-positive rapid plasma reagin test result showed a marked increase in relative risk for fetal loss, but the increase was not statistically significant because of the small number of women who tested positive for rapid plasma reagin. Our study is the first to consider the influence of confounding variables on the effect of aPL on pregnancy outcome. Our data on fetal loss are consistent with the conclusions of Lockwood and coworkers [21], who examined the risk of aPL positivity in normal pregnancy in a group of predominantly black women and found a seven- to eightfold increase in relative risk for fetal loss. This higher relative risk reported by these investigators compared with our findings may have resulted, in part, because those investigators did not enter confounding variables into their analysis. Nevertheless, it is striking that, despite the differences in ethnic composition of the two populations, the presence of aPL constituted a significant risk for fetal loss in both studies.

It is noteworthy that 24.4% of our pregnant women had an abnormal aPL profile, defined as an abnormal result on any single aPL test. A significant portion of the patients with aPL positivity in our study consisted of those with an abnormal dilute Russell viper venom time. Data on the lupus anticoagulant in healthy women is sparse, and the proportion of our participants with an abnormal lupus anticoagulant is higher than that previously reported for nonpregnant women [14]. Alterations in coagulation factors occur during pregnancy [28] that can affect the normal range of anticoagulant tests, such as the activated partial thromboplastin time [29]. The normal range of the dilute Russell viper venom time in this study was derived from healthy, nonsmoking, nonpregnant women and men who fasted. Although not previously shown [30], the use of frozen specimens may also have contributed to our high proportion of positive results. Thus, these factors may need to be considered when defining the "normal" range of dilute Russell viper venom time in pregnant women.

Our data indicate that only a few of our patients had more than one positive aPL test result and that aPL positivity varied considerably between the first prenatal visit and delivery. These observations are consistent with the work of Out and colleagues [31], who showed that some patients have persistently elevated levels of aPL and others have levels that fluctuate significantly over time. These investigators postulated that patients with persistently elevated levels of aPL are at the highest risk for complications. Unfortunately, our data do not allow us to address this question because repeated measures of aPL were not obtained for most of the patients who experienced early fetal loss. Nevertheless, our data emphasize that aPL levels may fluctuate during pregnancy. In contrast to previous reports [16, 26], the aPL level in the aPL-positive women with fetal loss was relatively low. However, in keeping with the findings in other patient groups [16] and recent animal studies [32], our results suggest that IgG anticardiolipin antibody is a better predictor of fetal loss than any other measure of aPL.

We postulated that aPL may act with several other risk factors to cause clinical symptoms. Therefore, we examined the effect of confounding variables on fetal outcome and entered these variables into logistic regression analysis with aPL. This analysis showed that a positive aPL profile and IgG anticardiolipin antibody were important risk factors for fetal loss independent of other variables. However, nearly 85% of patients with an abnormal aPL profile at the first prenatal visit had a normal pregnancy. The role of aPL in pregnancy loss is complex [21, 24]. The specific characteristics of the aPL and the presence of other factors that interact with aPL may play a critical role in determining the pathogenicity of aPL. Until these co-risk factors are identified [33] and pathogenic subsets of aPL [34] are better defined, screening of healthy pregnant women for the presence of aPL is not justified.

We could not show an association between aPL and maternal complications of pregnancy. This correlated with the findings of Harris and Spinnato [22] but was in contrast to other studies of selected [25, 35] and unselected patient groups [24] that showed that aPL positivity is associated with materno-fetal complications such as pregnancy-induced hypertension, preeclamptic toxemia, and low birth weight. However, complications in our study were few (thus reducing the power of the study), which may account for our inability to detect an increase in pregnancy complications in the aPL-positive group. Birth weights and Apgar scores of infants born to aPL-positive mothers were actually greater than these variables in infants born to aPL-negative mothers. Polzin and associates [25] showed an association between aPL positivity and intrauterine growth retardation based on examination of a highly select group of women. Our contrasting findings may have resulted from selection bias or may reflect a complex interaction of aPL positivity and intrauterine growth retardation; that is, intrauterine growth retardation in our aPL-positive mothers may have resulted in fetal death, and therefore only large fetuses could survive in aPL-positive mothers.

We showed that the presence of elevated levels of aPL at the first prenatal visit (< 25 weeks gestation) in women with normal pregnancies is associated with an increased risk for fetal loss. A positive aPL result was not associated with other complications of pregnancy, low live birth weights, or low Apgar scores. Comparison of aPL measurement at the first prenatal visit and at delivery showed variations in aPL levels, suggesting that aPL levels fluctuate during pregnancy and that serial studies of aPL may be necessary to better define the clinical associations of these antibodies. Despite the increased risk for fetal loss associated with aPL positivity, we showed clearly that in the absence of any therapeutic interventions, a subgroup of our aPL-positive cohort had successful pregnancy outcomes. Therefore, it is not appropriate to screen for aPL in first pregnancies until the indications for therapeutic intervention are more clearly defined and other factors that contribute to aPL pathogenicity [36-38] are better understood.