- Email: [email protected]
The effect of immunoglobulin G fractions from patients with lupus anticoagulant on placental prostacyclin and thromboxane production
The effect of immunoglobulin G fractions from patients with lupus anticoagulant on placental prostacyclin and thromboxane production Alan M. Peaceman, MD, and Karen A. Rehnberg, BA Chicago, Illinois OBJECTIVE: We evaluated whether the production of prostacyclin and thromboxane by normal human placental tissue is consistently altered by incubation with immunoglobulin G fractions prepared from plasma of patients with lupus anticoagulant. STUDY DESIGN: The immunoglobulin G fractions were prepared from eight patients with lupus anticoagulant and eight control patients. Doses of these fractions (3 mg, 7.5 mg, and 12 mg) were incubated with placental explants obtained from normal pregnancies, and prostacyclin and thromboxane production was assessed over 48 hours. RESULTS: Prostacyclin production was similar for placental tissue incubated with immunoglobulin fractions from control and lupus anticoagulant patients at all of the doses tested. Placental production of thromboxane was significantly increased with immunoglobulin fractions from lupus anticoagulant patients for all three doses (p = 0.02). CONCLUSIONS: The immunoglobulin G fraction from patients with lupus anticoagulant consistently alters placental thromboxane production without affecting prostacyclin production. Increases in placental thromboxane production may contribute to antiphospholipid antibody-mediated pregnancy loss. (AM J OBSTET GVNECOL 1993;169:1403-6.)
Key words: Antiphospholipid antibodies, lupus anticoagulant, thromboxane, prostacyclin, placenta Lupus anticoagulant is an antiphospholipid antibody that has been associated with poor obstetric histories, including recurrent abortion, fetal death, severe preeclampsia, and intrauterine fetal growth retardation. 1. 2 Initial data suggested that this antibody disrupts vascular endothelial cells, resulting in decreased prostacyclin production." The decrease in prostacyclin production could then result in clinical symptoms such as thrombosis. However, this finding of altered prostacyclin production by endothelial cells could not be consistently reproduced with plasma taken from all patients with antiphospholipid antibodies.v" As a result, the site and mechanism of action for antiphospholipid antibodies remain unproved. The choice of appropriate clinical treatments is hampered by the absence of a defined pathophysiologic mechanism for antiphospholipid antibody-mediated pregnancy loss. From the Section of Maternal-Fetal Medicine, Department ofObstetrics and Gynecology, Northwestern University Medical School and Northwestern Memorial Hospital. Presented at the Fortieth Annual Meeting of the Societyfor Gynecologic Investigation, Toronto, Ontario, Canada, March 31-April 3,
1993. Reprint requests: Alan M. Peaceman, MD, 333 E. Superior St., Suite 410, Chicago, IL 60611. Copyright © 1993 by Mosby-Year Book, Inc. 0002-9378/93 $1.00 + .20 6/6/51003
An alternate site for the action of antiphospholipid antibodies may be at the surface of the placenta rather than the vascular endothelium. Placental tissue is known to be active in producing prostaglandins, and an effect on placental prostanoid production could result in paracrine alteration of the local environment. Decreased prostacyclin and increased thromboxane production by the placenta has previously been associated with vasoconstriction, platelet aggregation, thrombosis, and decreased uteroplacental blood flow. These findings are pathologic features often associated with the antiphospholipid syndrome. It is therefore plausible that antiphospholipid antibodies are active at the maternal-placental interface and alter prostanoid production at that level. A previous investigation demonstrated that when the immunoglobulin G fraction from a single patient with lupus anticoagulant was incubated with tissue from multiple normal placentas, a consistent increase in thromboxane production was observed while prostacyclin production was unaffected." The current study evaluated the hypothesis that the IgG fractions from the plasma of multiple patients with lupus anticoagulant consistently alter the production of prostacyclin and thromboxane when incubated with normal human placental tissue.
1403
1404 Peaceman and Rehnberg
December 1993 Am J Obstet Gynecol
4000
3500
3000
E
~
2500 CONTROL 2000
0
c,
<3
1500
LA
u,
"u,
iooo 500 0 10
0
20
A
Time
30
4000
4000
3500
3500 3000
3000
E
~
E
2500
l
2000
.c
0
c,
1500
~
CONTROL LA
1000
2000 1500
u,
is'
CONTROL
1000
500
500
0
0 0
B
2500
<3
<3
"n.
so
40
(hours)
10
20 Time
30
40
0
50
(hours)
C
10
20 Time
30 (hours)
Fig. 1. Effect of IgG fractions from patients with lupus anticoagulant and controls on prostacyciin production by normal human placental tissue. Prostacyciin production over 48 hours for each dose of added IgG is assessed by measurement of the stable metabolite 6-keto-prostaglandin F 1« (PGF 1 alpha) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with IgG from patients with lupus anticoagulant (LA) are compared with those from tissue incubated with IgG from control patients. Three doses ofIgG tested were as follows: A, 3.0 mg; B, 7.5 mg; C, 12 mg. Data represent mean ± SE (n = 8).
Methods Whole blood was obtained by venipuncture in tubes containing sodium citrate from eight nonpregnant volunteers who served as controls. These samples were centrifuged at 3000g to remove blood cells and platelets, and the supernatant was centrifuged again at 3000g to ensure no cellular contamination. Each plasma sample obtained was treated with sufficient calcium chloride to achieve a final concentration of 5 mmol/L, and the resulting clot was removed with a glass rod with gentle stirring. This procedure enabled removal of the coagulation factors without release of platelet thromboxane. An equal volume of a saturated solution of ammonium sulfate (27.5%) was added to each sample, precipitating the IgG fractions. Each IgG fraction was resuspended in phosphate-buffered saline solution and dialyzed to remove traces of the ammonium sulfate. IgG fractions obtained in this manner contain other immunoglobulins and proteins, but enhanced IgG content has been confirmed previously by gel electrophoresis." Whole blood was obtained in a similar fashion from eight nonpregnant patients with demonstrable lupus
anticoagulant. These plasma samples were processed in the same manner as were the control samples, and ammonium sulfate precipitation of the IgG fractions was also performed. All control IgG and lupus anticoagulant IgG samples were then frozen and stored for later use. Fresh placentas were obtained from uncomplicated term pregnancies and processed within 15 minutes of delivery under sterile conditions. After removal of the basal plate, multiple placental explants of approximately 350 mg were dissected from the placenta, washed repeatedly with Dulbecco's modified Eagle medium (Gibco, Grand Island, N.Y.), and placed into separate incubation wells. To each well was added an aliquot of either a control or a lupus anticoagulant IgG sample, and a volume of 7 ml was achieved. Three doses of IgG were tested for each patient sample-3.0, 7.5, and 12 mg. These three doses of IgG produced concentrations in the wells corresponding to 4%, 10%, and 16% of maternal plasma levels. The wells were incubated at 37° C in a Dubnoff shaker, and aliquots were removed at 0, 8, 20, 32, and 48 hours. These aliquots were assessed for placental prostacyclin and
Peaceman and Rehnberg 1405
Volume 169, Number 6 Am J Obstet Gynecol
1000 900 800 700
E
600
~
SOD
N
400
!;!
f-
LA
300 I ;r---.f---i------~
CONTROL
200 100
10
A
20
30
Time
1000
E
ci; ~ N
m
50
1000
900
900
800
800
700
700
E
600 SOD
*
*
400
LA
~
300
600
N
co
400
200
100
100 40
50
(hours)
CONTROL
300
CONTROL
200
Time
LA
500
i"
i"
B
40
(hours)
10
C
20 TIME
30
40
50
(hours)
Fig. 2. Effect of IgG fractions from patients with lupus anticoagulant and controls on thromboxane production by normal human placental tissue. Thromboxane production over 48 hours for each dose of added IgG is assessed by measurement of the stable metabolite thromboxane B 2 (TXB2) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with IgG from patients with lupus anticoagulant (LA) are compared with those from tissue incubated with IgG from control patients. Three doses of IgG tested were as follows: A, 3.0 mg; B, 7.5 mg; C, 12 mg. Data represent mean ± SE (n = 8). Asterisk, p < 0.05 by repeated-measures testing, lupus anticoagulant IgG versus control.
thromboxane production by the radioimmunoassay of their stable metabolites, 6-keto-prostaglandin F 1", and thromboxane B 2 • 10 All test wells were run in duplicate with the results averaged for analysis. Data are expressed in picograms per milligram of wet tissue and displayed as mean ± 1 SE. Statistical differences in prostanoid production were analyzed with repeatedmeasures testing. This project was approved by the Institutional Review Board of Northwestern University Medical School. Results
Prostacyclin production by placental tissue incubated with the three doses of IgG from control patients and patients with lupus anticoagulant are compared in Fig. 1, A to C. Production rates did not differ between these two patient groups with any of the doses tested. In addition, increasing doses of IgG fractions were not associated with differences in prostacyclin production within either control or lupus anticoagulant patients.
Thromboxane production by placental tissue incubated with 3 mg of IgG from control patients and patients with lupus anticoagulant is displayed in Fig. 2, A. On comparison with controls, addition of IgG from lupus anticoagulant-positive patients was associated with an increase in thromboxane production by 20 hours of incubation (p < 0.05). The addition of IgG fractions from all eight lupus anticoagulant patients was associated with at least one value for thromboxane production above the 95% confidence limit for thromboxane production from wells containing control IgG. Similar increases in thromboxane production were also noted for placental explants incubated with IgG fractions from patients with lupus anticoagulant at the two higher doses tested (Fig. 2, B and C). For both 7.5 and 12 mg of added IgG, this increase was statistically significant by 20 hours of incubation (p < 0.05). In addition, thromboxane production with 12 mg of added lupus anticoagulant IgG was significantly higher than with 3 mg of added lupus anticoagulant IgG
1406 Peaceman and Rehnberg
(P < 0.05). In contrast to the uniform pattern observed with 3 mg of added lupus anticoagulant IgG, two lupus anticoagulant IgG samples tested at 7.5 mg and one sample tested at 12 mg did not have any thromboxane levels above the 95% confidence limits established from control wells. Comment The addition of IgG fractions from patients with lupus anticoagulant to placental tissue explants resulted in significantly higher thromboxane levels compared with controls at all doses tested. Although thromboxane levels with the 3 mg and 7.5 mg doses were similar and variation among patients was significant, the increase with 12 mg suggests a dose-response relationship between thromboxane production and added lupus anticoagulant IgG. In contrast, thromboxane production with the higher doses of control IgG was unaltered from levels seen with 3 mg of added control IgG and similar to previously reported levels produced by placental explants without added plasma fractions." Also, no alteration in placental prostacyclin production was seen with the addition of either control or lupus anticoagulant IgG fractions at any concentration tested. These results confirm our previous observations in which the IgG fraction from a single patient with lupus anticoagulant was incubated with placental tissue obtained from multiple normal pregnancies." Carreras et al." initially suggested that lupus anticoagulant caused clinical disease by interacting with vascular endothelial cells, with a resultant decrease in prostacyclin production. Our data support an alternative mechanism of action for antiphospholipid antibodies - an increase in thromboxane production by the placenta. Increased platelet thromboxane production has been associated with thrombosis, a major clinical and pathologic finding in patients with lupus anticoagulant. II The inability of previous investigators to find an alteration in thromboxane production when antiphospholipid antibodies were incubated with endothelial cells" may be explained by the placenta's being the primary site of action. It has not been established that thromboxane produced by the placenta can act system-
December 1993 Am J Obstet Gynecol
ically in causing thrombosis; however, it may contribute to thrombosis at the uteroplacental interface. These data support the hypothesis that antiphospholipid antibodies have a direct effect on the placenta, which contributes to adverse pregnancy outcome. An alteration in placental prostanoid production, specifically thromboxane rather than prostacyclin production, may mediate the pathophysiologic process. Further investigations will be needed to confirm this site of action and to identify which cellular components of the placenta are the specific targets. REFERENCES 1. Reece EA, Gabrielli S, Cullen MT, Zheng XZ, Hobbins JC, Harris EN. Recurrent adverse pregnancy outcome and antiphospholipid antibodies. AM 1 OBSTET GYNECOL 1990; 163:162-9. 2. Branch DW, Andres R, Digre KB, Rote NS, Scott JR. The association of anti phospholipid antibodies with severe preeclampsia. Obstet Gynecol 1989;73:541-5. 3. Carreras LO, Machin SJ, Deman R, et al. Arterial thrombosis, intrauterine death and "lupus" anticoagulant: detection of immunoglobulin interfering with prostacyclin formation. Lancet 1981; 1:244-6. 4. Carreras LO, Vermylen lG. "Lupus" anticoagulant and thrombosis - possible role of inhibition of prostacyclin formation. Thromb Haemost 1982;48:38-40. 5. Ros 10, Tarres MY, Baucells MY, Maired 11, Solano JT. Prednisone and maternal lupus anticoagulant. Lancet 1983;2:576. 6. Dudley D], Mitchell MD, Branch DW. Pathophysiology of antiphospholipid antibodies: absence of prostaglandinmediated effects on cultured endothelium. AM J OBSTIT GYNECOL 1990; 162:953-9. 7. Cariou R, Tobelem G, Belluci S, et al. Effect of lupus anticoagulant on anti thrombogenic properties of endothelial cells-inhibition of thrombomodulin-dependent protein C activation. Thromb Haemost 1988;60:54-8. 8. Walker TS, Triplett DA, 1aved N, Musgrave K. Evaluation of lupus anticoagulant: antiphospholipid antibodies, endothelium associated immunoglobulin, endothelial prostacyclin secretion, and antigenic protein S levels. Thromb Res 1988;51:267-81. 9. Peaceman AM, Rehnberg K. The immunoglobulin G fraction from plasma containing antiphospholipid antibodies causes increased placental thromboxane production. AM 1 OBSTIT GYNECOL 1992;167:1543-7. 10. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J OBSTET GYNECOL 1985;152:335-40. 11. Scott JR, Rote NS, Branch DW. Immunologic aspects of recurrent abortion and fetal death. Obstet Gynecol 1987; 70:645-56.
Key words: Antiphospholipid antibodies, lupus anticoagulant, thromboxane, prostacyclin, placenta Lupus anticoagulant is an antiphospholipid antibody that has been associated with poor obstetric histories, including recurrent abortion, fetal death, severe preeclampsia, and intrauterine fetal growth retardation. 1. 2 Initial data suggested that this antibody disrupts vascular endothelial cells, resulting in decreased prostacyclin production." The decrease in prostacyclin production could then result in clinical symptoms such as thrombosis. However, this finding of altered prostacyclin production by endothelial cells could not be consistently reproduced with plasma taken from all patients with antiphospholipid antibodies.v" As a result, the site and mechanism of action for antiphospholipid antibodies remain unproved. The choice of appropriate clinical treatments is hampered by the absence of a defined pathophysiologic mechanism for antiphospholipid antibody-mediated pregnancy loss. From the Section of Maternal-Fetal Medicine, Department ofObstetrics and Gynecology, Northwestern University Medical School and Northwestern Memorial Hospital. Presented at the Fortieth Annual Meeting of the Societyfor Gynecologic Investigation, Toronto, Ontario, Canada, March 31-April 3,
1993. Reprint requests: Alan M. Peaceman, MD, 333 E. Superior St., Suite 410, Chicago, IL 60611. Copyright © 1993 by Mosby-Year Book, Inc. 0002-9378/93 $1.00 + .20 6/6/51003
An alternate site for the action of antiphospholipid antibodies may be at the surface of the placenta rather than the vascular endothelium. Placental tissue is known to be active in producing prostaglandins, and an effect on placental prostanoid production could result in paracrine alteration of the local environment. Decreased prostacyclin and increased thromboxane production by the placenta has previously been associated with vasoconstriction, platelet aggregation, thrombosis, and decreased uteroplacental blood flow. These findings are pathologic features often associated with the antiphospholipid syndrome. It is therefore plausible that antiphospholipid antibodies are active at the maternal-placental interface and alter prostanoid production at that level. A previous investigation demonstrated that when the immunoglobulin G fraction from a single patient with lupus anticoagulant was incubated with tissue from multiple normal placentas, a consistent increase in thromboxane production was observed while prostacyclin production was unaffected." The current study evaluated the hypothesis that the IgG fractions from the plasma of multiple patients with lupus anticoagulant consistently alter the production of prostacyclin and thromboxane when incubated with normal human placental tissue.
1403
1404 Peaceman and Rehnberg
December 1993 Am J Obstet Gynecol
4000
3500
3000
E
~
2500 CONTROL 2000
0
c,
<3
1500
LA
u,
"u,
iooo 500 0 10
0
20
A
Time
30
4000
4000
3500
3500 3000
3000
E
~
E
2500
l
2000
.c
0
c,
1500
~
CONTROL LA
1000
2000 1500
u,
is'
CONTROL
1000
500
500
0
0 0
B
2500
<3
<3
"n.
so
40
(hours)
10
20 Time
30
40
0
50
(hours)
C
10
20 Time
30 (hours)
Fig. 1. Effect of IgG fractions from patients with lupus anticoagulant and controls on prostacyciin production by normal human placental tissue. Prostacyciin production over 48 hours for each dose of added IgG is assessed by measurement of the stable metabolite 6-keto-prostaglandin F 1« (PGF 1 alpha) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with IgG from patients with lupus anticoagulant (LA) are compared with those from tissue incubated with IgG from control patients. Three doses ofIgG tested were as follows: A, 3.0 mg; B, 7.5 mg; C, 12 mg. Data represent mean ± SE (n = 8).
Methods Whole blood was obtained by venipuncture in tubes containing sodium citrate from eight nonpregnant volunteers who served as controls. These samples were centrifuged at 3000g to remove blood cells and platelets, and the supernatant was centrifuged again at 3000g to ensure no cellular contamination. Each plasma sample obtained was treated with sufficient calcium chloride to achieve a final concentration of 5 mmol/L, and the resulting clot was removed with a glass rod with gentle stirring. This procedure enabled removal of the coagulation factors without release of platelet thromboxane. An equal volume of a saturated solution of ammonium sulfate (27.5%) was added to each sample, precipitating the IgG fractions. Each IgG fraction was resuspended in phosphate-buffered saline solution and dialyzed to remove traces of the ammonium sulfate. IgG fractions obtained in this manner contain other immunoglobulins and proteins, but enhanced IgG content has been confirmed previously by gel electrophoresis." Whole blood was obtained in a similar fashion from eight nonpregnant patients with demonstrable lupus
anticoagulant. These plasma samples were processed in the same manner as were the control samples, and ammonium sulfate precipitation of the IgG fractions was also performed. All control IgG and lupus anticoagulant IgG samples were then frozen and stored for later use. Fresh placentas were obtained from uncomplicated term pregnancies and processed within 15 minutes of delivery under sterile conditions. After removal of the basal plate, multiple placental explants of approximately 350 mg were dissected from the placenta, washed repeatedly with Dulbecco's modified Eagle medium (Gibco, Grand Island, N.Y.), and placed into separate incubation wells. To each well was added an aliquot of either a control or a lupus anticoagulant IgG sample, and a volume of 7 ml was achieved. Three doses of IgG were tested for each patient sample-3.0, 7.5, and 12 mg. These three doses of IgG produced concentrations in the wells corresponding to 4%, 10%, and 16% of maternal plasma levels. The wells were incubated at 37° C in a Dubnoff shaker, and aliquots were removed at 0, 8, 20, 32, and 48 hours. These aliquots were assessed for placental prostacyclin and
Peaceman and Rehnberg 1405
Volume 169, Number 6 Am J Obstet Gynecol
1000 900 800 700
E
600
~
SOD
N
400
!;!
f-
LA
300 I ;r---.f---i------~
CONTROL
200 100
10
A
20
30
Time
1000
E
ci; ~ N
m
50
1000
900
900
800
800
700
700
E
600 SOD
*
*
400
LA
~
300
600
N
co
400
200
100
100 40
50
(hours)
CONTROL
300
CONTROL
200
Time
LA
500
i"
i"
B
40
(hours)
10
C
20 TIME
30
40
50
(hours)
Fig. 2. Effect of IgG fractions from patients with lupus anticoagulant and controls on thromboxane production by normal human placental tissue. Thromboxane production over 48 hours for each dose of added IgG is assessed by measurement of the stable metabolite thromboxane B 2 (TXB2) and is expressed in picograms per milligram of wet tissue. Measurements obtained from tissue incubated with IgG from patients with lupus anticoagulant (LA) are compared with those from tissue incubated with IgG from control patients. Three doses of IgG tested were as follows: A, 3.0 mg; B, 7.5 mg; C, 12 mg. Data represent mean ± SE (n = 8). Asterisk, p < 0.05 by repeated-measures testing, lupus anticoagulant IgG versus control.
thromboxane production by the radioimmunoassay of their stable metabolites, 6-keto-prostaglandin F 1", and thromboxane B 2 • 10 All test wells were run in duplicate with the results averaged for analysis. Data are expressed in picograms per milligram of wet tissue and displayed as mean ± 1 SE. Statistical differences in prostanoid production were analyzed with repeatedmeasures testing. This project was approved by the Institutional Review Board of Northwestern University Medical School. Results
Prostacyclin production by placental tissue incubated with the three doses of IgG from control patients and patients with lupus anticoagulant are compared in Fig. 1, A to C. Production rates did not differ between these two patient groups with any of the doses tested. In addition, increasing doses of IgG fractions were not associated with differences in prostacyclin production within either control or lupus anticoagulant patients.
Thromboxane production by placental tissue incubated with 3 mg of IgG from control patients and patients with lupus anticoagulant is displayed in Fig. 2, A. On comparison with controls, addition of IgG from lupus anticoagulant-positive patients was associated with an increase in thromboxane production by 20 hours of incubation (p < 0.05). The addition of IgG fractions from all eight lupus anticoagulant patients was associated with at least one value for thromboxane production above the 95% confidence limit for thromboxane production from wells containing control IgG. Similar increases in thromboxane production were also noted for placental explants incubated with IgG fractions from patients with lupus anticoagulant at the two higher doses tested (Fig. 2, B and C). For both 7.5 and 12 mg of added IgG, this increase was statistically significant by 20 hours of incubation (p < 0.05). In addition, thromboxane production with 12 mg of added lupus anticoagulant IgG was significantly higher than with 3 mg of added lupus anticoagulant IgG
1406 Peaceman and Rehnberg
(P < 0.05). In contrast to the uniform pattern observed with 3 mg of added lupus anticoagulant IgG, two lupus anticoagulant IgG samples tested at 7.5 mg and one sample tested at 12 mg did not have any thromboxane levels above the 95% confidence limits established from control wells. Comment The addition of IgG fractions from patients with lupus anticoagulant to placental tissue explants resulted in significantly higher thromboxane levels compared with controls at all doses tested. Although thromboxane levels with the 3 mg and 7.5 mg doses were similar and variation among patients was significant, the increase with 12 mg suggests a dose-response relationship between thromboxane production and added lupus anticoagulant IgG. In contrast, thromboxane production with the higher doses of control IgG was unaltered from levels seen with 3 mg of added control IgG and similar to previously reported levels produced by placental explants without added plasma fractions." Also, no alteration in placental prostacyclin production was seen with the addition of either control or lupus anticoagulant IgG fractions at any concentration tested. These results confirm our previous observations in which the IgG fraction from a single patient with lupus anticoagulant was incubated with placental tissue obtained from multiple normal pregnancies." Carreras et al." initially suggested that lupus anticoagulant caused clinical disease by interacting with vascular endothelial cells, with a resultant decrease in prostacyclin production. Our data support an alternative mechanism of action for antiphospholipid antibodies - an increase in thromboxane production by the placenta. Increased platelet thromboxane production has been associated with thrombosis, a major clinical and pathologic finding in patients with lupus anticoagulant. II The inability of previous investigators to find an alteration in thromboxane production when antiphospholipid antibodies were incubated with endothelial cells" may be explained by the placenta's being the primary site of action. It has not been established that thromboxane produced by the placenta can act system-
December 1993 Am J Obstet Gynecol
ically in causing thrombosis; however, it may contribute to thrombosis at the uteroplacental interface. These data support the hypothesis that antiphospholipid antibodies have a direct effect on the placenta, which contributes to adverse pregnancy outcome. An alteration in placental prostanoid production, specifically thromboxane rather than prostacyclin production, may mediate the pathophysiologic process. Further investigations will be needed to confirm this site of action and to identify which cellular components of the placenta are the specific targets. REFERENCES 1. Reece EA, Gabrielli S, Cullen MT, Zheng XZ, Hobbins JC, Harris EN. Recurrent adverse pregnancy outcome and antiphospholipid antibodies. AM 1 OBSTET GYNECOL 1990; 163:162-9. 2. Branch DW, Andres R, Digre KB, Rote NS, Scott JR. The association of anti phospholipid antibodies with severe preeclampsia. Obstet Gynecol 1989;73:541-5. 3. Carreras LO, Machin SJ, Deman R, et al. Arterial thrombosis, intrauterine death and "lupus" anticoagulant: detection of immunoglobulin interfering with prostacyclin formation. Lancet 1981; 1:244-6. 4. Carreras LO, Vermylen lG. "Lupus" anticoagulant and thrombosis - possible role of inhibition of prostacyclin formation. Thromb Haemost 1982;48:38-40. 5. Ros 10, Tarres MY, Baucells MY, Maired 11, Solano JT. Prednisone and maternal lupus anticoagulant. Lancet 1983;2:576. 6. Dudley D], Mitchell MD, Branch DW. Pathophysiology of antiphospholipid antibodies: absence of prostaglandinmediated effects on cultured endothelium. AM J OBSTIT GYNECOL 1990; 162:953-9. 7. Cariou R, Tobelem G, Belluci S, et al. Effect of lupus anticoagulant on anti thrombogenic properties of endothelial cells-inhibition of thrombomodulin-dependent protein C activation. Thromb Haemost 1988;60:54-8. 8. Walker TS, Triplett DA, 1aved N, Musgrave K. Evaluation of lupus anticoagulant: antiphospholipid antibodies, endothelium associated immunoglobulin, endothelial prostacyclin secretion, and antigenic protein S levels. Thromb Res 1988;51:267-81. 9. Peaceman AM, Rehnberg K. The immunoglobulin G fraction from plasma containing antiphospholipid antibodies causes increased placental thromboxane production. AM 1 OBSTIT GYNECOL 1992;167:1543-7. 10. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J OBSTET GYNECOL 1985;152:335-40. 11. Scott JR, Rote NS, Branch DW. Immunologic aspects of recurrent abortion and fetal death. Obstet Gynecol 1987; 70:645-56.