AMERICAN JOURNAL OF REPRODUCTIVE IMMUNOLOGY 28:199-204 © 1992 MUNKSGAARD
The Effect of Systemic Lupus Erythematosus on Pregnancy and Pregnancy Outcome JOHN P. HAYSLETT Department ofMedicine, Section ofNephrology, Yale University School of Medicine, New Haven, Connecticut ABSTRACT: This review provides an analysis of reports published since 1980 on the effect of systemic lupus erythematosus (SLE) on pregnancy and pregnancy outcome. The question whether pregnancy increases clinical flares and the severity of flares in patients with SLE during pregnancy has not been resolved because of difficulty in defining exacerbations of SLE and of preeclampsia. An analysis of major detailed reports indicates that maternal complications are reduced in patients who are in clinical remission prior to the onset of pregnancy compared with women with persistent disease activity. Complications are observed in 30%-50% of patients with inactive disease at onset of gestation. After exclusion of spontaneous abortions during the first trimester, fetal survival was 85%-90% in most reported case series. The best outcomes were reported in patients with inactive disease at onset of pregnancy. It seems likely that some maternal complications and fetal wastage in this population are related to anticardiolipin antibodies. (Am J Reprod Immunol. 1992; 28:199-204.) Key words: Anticardiolipin antibodies, fetal outcome, maternal complications, pregnancy-induced exacerbations. INTRODUCTION
Systemic lupus erythematosus (SLE) is a disease that primarily affects women of child-bearing age. Whether SLE is affected by pregnancy and whether the likelihood for a successful outcome is reduced in patients with SLE are issues of concern. The scientific basis for answering these questions has improved recently because of numerous reports on relatively large series of patients with SLE who have experienced pregnancy. This review analyzes the larger series published since 1980, especially those that include detailed information. It should be recalled that the clinical literature related to SLE and pregnancy is confounded by wide differences in patient populations among individual reports. Some reports include cases with severe forms ofSLE, while others involve women with milder disease. In addition, there are series that include patients selected from multiple centers, as well as consecutive series of cases observed in a single clinic. In most reports, the quality of obstetrical care and neonatal management is unknown, and treatment regimens for SLE are not described. Nevertheless, these reports provide the database that is available for analysis.
Submitted for publication September 1, 1992; accepted September 8, 1992. Address reprint requests to John P. Hayslett, M.D., Department of Medicine, Yale School of Medicine, 333 Cedar Street - LMP 2073, New Haven, CT 06510-8056.
EXACERBATION OF SLE DURING PREGNANCY
The question whether pregnancy adversely affects SLE and causes a heightened incidence of clinical flares and/or increased severity of flares remains undetermined. On a theoretical basis an adverse effect on the underlying disease diathesis would be expected. There is evidence that estrogens have a potent influence on the immune system and on the initiation of SLE in man and in animal models ofSLE. The female:male ratio in man is 9:1 and SLE is reported in men with Klinefelter's syndrome. 1 Moreover, in the NZB x NZW F1 murine model ofSLE, the disease occurs predominantly in females but can be produced in males by administration of estrogens. 2 Sex hormones modulate the expression of autoimmunity and affect the concentration of antibodies to nucleic acid, the severity of glomerulonephritis, and, ultimately, mortality. In two retrospective clinical studes3.4 a marked increase in clinical exacerbations was observed during pregnancy and the puerperium, compared with the apparent incidence in the same patients prior to pregnancy. In a;rospective analysis, however, Lockshin and associates compared 21 patients with mild SLE who attempted to complete pregnancy with nonpregnant women with SLE matched for age, race, and severity of disease. The criteria used for clinical flares excluded new onset of proteinuria (four cases), regarded as probable preeclampsia, and new onset thrombocytopenia (eight cases in the pregnancy group compared with one in control patients). These investigators concluded that there was no difference in the incidence of exacerbations or in disease activity between pregnant and nonpregnant patients. Ifthese exclusive criteria for defining flares ofSLE had not been used in analysis, a marked increase in disease activity during pregnancy would have been recorded. A subsequent report by Mintz and associates" compared disease activity in a larger patient series of 102 pregnancies in 75 patients with a concurrent series of patients who did not experience pregnancy, and also found no difference in the apparent incidence of exacerbation. Details of these populations were not provided in their report. An explanation of the different results in these studies is confounded by the difficulty in establishing criteria for clinical flares of SLE, the absence of a firm clinical marker for preeclampsia, and the issue whether maternal complications thought to be caused by anticardiolipin antibodies should be regarded as flares of SLE. An attempt to resolve this question will be an important challenge for future studies. Regardless of the outcome, the reported incidence of exacerbations of 10%-75% during pregnancy and postpartum and the high fetal wastage of up to 40% in some reports highlight the concern for the potential of this group of patients to bear children successfully. In the series reported by Lockshin" with mild disease there was a 44% incidence of preterm deliveries and a fetal survival of only 68%.
200
HAYSLETT
ESTABLISHED SLE - INCIDENCE OF EXACERBATION REMISSION AT ONSET
ACTIVITY AT ONSET
100
100
80
80
60
60
40
40
20
20
o
o
o
Hayslett, Kid Inti 18:207, 1980
ELZI
Bobrie, Am J Kid Dis 9:339, 1987 ILl Houser, Am J OB/GYN 138:409, 1980 Fig.1.
e
Tozman, J Rheum 7:624, 1980
§
Lockshin, Arth & Rheum 32:665, 1989
Correlations of relapse or exacerbations ofSLE in relationship to activity of disease at onset of pregnancy.
MATERNAL COMPLICATIONS DURING PREGNANCY AND POSTPARTUM
In 1965 Estes and Larson 7 suggested, in an analysis of79 pregnancies after onset of disease in 36 women, that the effect of pregnancy on SLE correlated directly with the activity of disease at the time of conception. It is of interest that in this report, which spanned 20 years, glucocorticoid treatment was not available in more than onehalf of patients, and therefore the series may have included women with milder forms of SLE. In 35 pregnancies in which SLE was in remission at the start of pregnancy, relapse occurred in 26%, while remission was sustained in 74%. In contrast, SLE remained active in 37% and exacerbation occurred in 33% of pregnancies associated with active disease at conception. This correlation is supported by reports since 1980. Results of five studies are shown in Fig. 1 in which SLE predated pregnancy and in which sufficient data were available to assess clinical activity of SLE at onset of pregnancy. Hayslett and Lynn" reported on 65 selected pregnancies by 47 patients. These cases were selected by nephrologists and were managed in multiple centers in North America. In 80% of cases clinical renal disease was present before the index pregnancy. Following complete clinical remission ofSLE for at least 6 months prior to conception there was sustained remission in 65%, while 32% were complicated by relapse during pregnancy. Complete or partial remission occurred in all cases within 3 months of delivery. In contrast, the course of pregnancy associated with disease activity at conception was marked
by persistent or increased activity of SLE in 52%. Exacerbations were more severe in this group and clinical remission did not occur after delivery in four cases; one patient died within 6 months of delivery. Bobrie and associates" evaluated 53 pregnancies in 35 women after the onset of SLE. These cases were also a selected series and were initially managed in multiple medical centers in an urban center in Europe. Exacerbations were observed in 62% of women with signs of active disease at onset of pregnancy, compared with 34% in cases with remission at onset of pregnancy. In the former "group clinical flares were more severe, and five patients progressed rapidly to end-stage renal disease, all of whom had severe extrarenal manifestations. Houser and associatea'? described the clinical course of 18 pregnancies in 11 individuals with onset ofSLE in childhood. All cases were managed in a single medical center. Of 10 pregnancies associated with complete remission at conception, gestation was uncomplicated in 80%, while in the remaining eight pregnancies associated with clinical activity, increased clinical manifestations or worsening of renal disease occurred in 50%; none, however, exhibited rapidly progressive renal failure during follow-ur, which averaged 4 years. Tozman and associates 1 reported the course of 24 pregnancies in 18 women followed by one rheumatology clinic. Of 11 instances with inactive disease at onset of pregnancy only one exhibited relapse. In contrast, the disease remained active throughout pregnancy in 85% of pregnancies marked by clinical activity at onset. Lockshin'"
PREGNANCY AND SLE
described the course of 80 patients with SLE. Although clinical activity at conception was not reported, I assume, for the purpose of analysis, that 53 patients, or 66%, had inactive disease as they were not treated with corticosteroid, while 27 patients on corticosteroid had active ~is ease. As shown in Fig. 1, the recurrence or exacerbation rates were 37% in cases with activity at onset of pregnancy and 19% in those with inactive disease. In this computation one patient with onset of disease during pregnancy and two patients with recurrence of rash ~r termination of treatment were excluded from analysis. The report of Mintz and associates" was not included in the composite analysis as only 10% of patients had active disease at conception, and in most of these severity appeared to be mild. This large series with inactive SLE at onset of pregnancy, managed by one medical center, provides important insights into the incidence and types of recurrences of lupus activity in patients with inactive disease at the onset of pregnancy. Of the 92 pregnancies that began while in remission, 55 (60%) had an exacerbation during pregnancy, postpartum, or postabortion. In most cases the manifestations of SLE were mild and involved the cutaneous or musculoskeletal systems. Hematological manifestations of leukopenia and thrombocytopenia occurred in 35%, exacerbation or new onset of nephropathy was observed in 16%, and 15% exhibited central nervous system (CNS) disease. There were no maternal deaths. When the onset of SLE occurs during pregnancy or postpartum, clinical manifestations tend to be severe. Of nine patients in the report by Hayslett and Lynn," five had evidence of nephrotic syndrome, with renal insufficiency in two, and all had severe nonrenal manifestations. Similar hectic clinical courses in pregnancies associated with initial onset of SLE were described by Bobrie and associates" and Imbasciati and associates.l" Taken together these six publications describe SLE activity during pregnancy in over 300 gestations. Although caution should be exercised in extrapolating from these reports, as they represent heterogeneous populations of patients, these results suggest, as reported by Estes and Larson7 30 years ago, that the frequency and severity of recurrences or exacerbations during pregnancy tend to be more common and more severe when pregnancy commences in the setting of active SLE. In addition, they indicate that even in patients with inactive disease there is a high incidence of recurrent disease, including manifestations that involve the kidney and CNS.
201
FETAL OUTCOME
Inspection of published reports prior to 1980 indicates a live birth rate by women with established SLE of approximately 60%, after exclusion of therapeutic abortions.l" In most of these reports, as well as in recent publications, all spontaneous abortions were listed. Inclusion of spontaneous abortions during the first trimester inflates the incidence offetal death in SLE-related pregnancies because the spontaneous rate in normal women is over 20% in the first trimester. 15 Table I lists data on fetal outcome in 302 pre§nancies from nine reports published since 1980. 4,6,8-1 .13, 6,17 Publications from the same era that failed to provide sufficient detail on fetal outcome were omitted from analysis. The number of pregnancies listed for each report reflects only pregnancies in women with established SLE before the index pregnancies and in which the fetus survived beyond 12 weeks. The incidence of clinically evident lupus nephropathy, which was 60% or more in seven reports, is listed for each series because its presence correlates with the severity of SLE and because the adverse effects of renal disease and SLE on fetal outcome are probably additive. The incidence oflate abortions, stillbirths, and neonatal deaths have been combined for convenience. Inspection of these data indicates a wide range of preterm deliveries, defined as delivery before 37 weeks' gestation. In general, preterm deliveries were more common in patients with more active SLE, but there is no explanation for the wide variation between different studies, except for the possibility of different attitudes toward delivery by Caesarian section or induction oflabor among different medical centers. In eight ofthe nine reports fetal survival exceeded 80%, despite the high incidence of lupus nephritis and often hectic maternal courses described in most of these studies. The reason for the low survival rate of 56% in the Imbasciati et al. 13 series is not clear, but may relate to referred patients with more severe systemic disease; severe renal insufficiency was reported in only one patient. Not suprisingly, when fetal outcome was correlated with maternal SLE activity, survival rate was higher in patients with a lower incidence of recurrence or exacerbation, as shown in Fig. 2. These data were abstracted from four reports8-11 of patients with established SLE prior to conception, where sufficient detail was provided to make such a correlation. In the group with inactive
TABLE I. Pregnancy Outcome in Patients With Established SLE*
Author
Hayslett" Bobrie? Tozman!' Houser!" Zulman" Mintz" Varner 16 Imbasciati 13 Wong l ?
No. pregnancies
% Renal
% Preterm
disease
births
% Deaths
% Fetal andneontal survival
47 38 18 14 24 102 26 16 17
80 100 61 100 67 59 21 100 12
4 13 8 21
13 13 11 7 4 11 0 38 0
81 87 89 93 83 89 100 56 100
"Deaths include second and third trimester abortions, stillbirths, and neonetal deaths.
49 12 69 41
202
HAYSLETT
ESTABLISHED SLE - FETAL SURVIVAL REMISSION AT ONSET
ACTIVITY AT ONSET
100 80
'E 60 Q) ~
Q)
a.. 40 20
o
Hayslett, Kidney Inter 18:207,1980
~ Dobrie, Am J Kid Dis 9:339, 1987
. . Houser, Am J Obstet Gynec 138: 409, 1980 ~ Tozman, J Rheum 7:624, 1980
Fig. 2. Fetal survival in patients with SLE in relationship to activity of disease at onset of pregnancy.
disease at conception, and therefore fewer flares during pregnancy, the live birth rate was 88%-100%. In contrast, in women with active disease at onset of gestation the live birth ranged between 50% and 82% in the same series. Maternal disease activity is usually reported to be severe when SLE initially began during pregnancy or immediately postpartum. In the three reports 8,9,16 that provided information on this subgroup, fetal survival was reduced to 50%-65%, as shown in Fig. 3. The effect of maternal SLE on fetal growth and development is not characterized solely by survival rate. The data in Table II were derived from the large series of Mintz and associates" managed by one center with modern perinatal care facilities. They demonstrate high rates of preterm deliveries and intrauterine growth impairment even in women with predominantly inactive disease at onset of pregnancy. Not surprisingly, the incidence of small-forgestational-age infants was higher in preterm deliveries and averaged 23% in the series as a whole. ROLE OF ANTICARDIOLIPIN ANTIBODIES IN MATERNAL COMPLICATIONS AND FETAL WASTAGE
Recent studies have suggested that antiphospholipid antibodies, including anticardiolipin and the lupus anticoagulant, play an important role in the pathogenesis
of vascular lesions in patients with SLE. 18,19 The average prevalence of anticardiolipin antibody is reported to be 44% in SLE, and that of the lupus anticoagulant, 34%.18 There is strong evidence that high antibody titers correlate with arterial and venous thrombosis, hemolytic anemia, thrombocytopenia, and recurrent fetal 10ss.20 Both types of antibodies recognize similar antigenic determinants, and are postulated to enhance platelet clearance by binding to membrane phospholipids or by promoting platelet activation.l" Recent evidence indicates that anticardiolipin enhances platelet-activating factor by vascular endothelial cells. 21 These types of mechanisms are probably responsible for the formation of platelet thrombi on vessel walls. In addition, activated platelets have a critical procoagulant role and serve as a nidus for membrane-based coagulation reactions involving the generation of thrombin and the formation of fibrin. It is thought that spontaneous abortion is caused by venous thrombosis in the placenta. The data, shown in Table III were derived from the report by Deleze and associates.~2In Group 1, comprised of patients with SLE and recurrent loss of three or more products of conception, nearly 80% of pregnancies were unsuccessful compared with patients with SLE who did not have a history of recurrent unsuccessful pregnancies (Group 2). The incidence of anticardiolipin was 83% in
PREGNANCY AND SLE
203
ONSET OF SLE DURING PREGNANCY OR POST PARTUM - FETAL SURVIVAL 100 80
C (l)
60
~
(l)
o,
40
20
o
Hayslett, Kidney Inter 18:207,1980
~ Dobrie, Am J Kid Dis 9:339, 1987
G Fig. 3.
Varner, Am J Obstet Gynec 145:1025, 1983
Fetal survival in patients with SLE who experienced the onset of disease during pregnancy or postpartum,
Group 1 and 14% in Group 2. The groups with SLE are contrasted with healthy patients with (Group 3) and without (Group 4) recurrent fetal wastage. In this study the likelihood of fetal loss in patients with SLE and anticardiolipin antibodies was approximately 10 times greater than in patients without antibody titers. In nearly all cases with fetal wastage the antibody was the IgG isoform. In another study of84 patients without a ~revi ous history of repeated spontaneous abortions," IgG anticardiolipin antibodies correlated with a spontaneous abortion rate of 59% in 46 patients, compared with the rate of 25% in 38 patients without the antibody. In addition, abortion occurred later in gestation (17.4 ± 7.1 wk) than in patients without antibodies (12.5 ± 7.1 wk). There is no evidence that treatment with glucocorticoids or antiplatelet agents, such as aspirin, reduces the incidence of fetal loss due to anticardiolipin antibodies. Because high titers of anticardiolipin antibodies have been associated with vascular and hematological complications in nonpregnant patients with SLE, it seems likely that some maternal complications during pregnancy are also caused by this mechanism. In the series reported by Deleze and associates.P there were statistically higher rates of thrombocytopenia, venous throm-
bosis, and convulsions in patients with anticardiolipin antibodies and recurrent fetal loss compared with cases without antibodies. In addition, recent studies suggest that antibody-related thrombosis may also playa role in exacerbations of renal disease during pregnancy. KincaidSmith and associates'" described the course of pregnancy in 12 women with circulating antiphospholipid antibodies, four of whom had SLE. Renal biopsy showed fibrin thrombi in glomeruli and arteries. Renal function was reduced in six of these patients and only two of23 pregnancies were successful. Although it seems clear that anticardiolipin antibodies play an important role in the mechanism offetal wastage and maternal complications in SLE, the relative importance of this factor has not been defined. There is, for example, a wide discrepancy between the rate offetal loss predicted from the reported prevalence of anticardiolipin antibodies in SLE of about 40% and an associated spontaneous abortion rate of 50%-60%, and the observed incidence of late abortions and stillbirths in series ofpregnant women with SLE oflO%-15%, reported before discovery of anticardiolipin antibodies. Clearly future studies are needed that include well characterized unselected patients; serial measurements of anticardiolipin before and during gestation; and controls matched for age, disease, and clinical history.
TABLE II. The Effect of SLE on Fetal Outcome* Characteristic
ActiveSLE
Inactive SLE
No. pregnancies
51
% Preterm deliveries % Abortions % Stillbirths % Small for gestational age
59
51 39
14
20
% Term deliveries
27
'Data from Mintz et al. 6
3
2
23
41
CONCLUSION
In summary, important strides have been made during the past decade in understanding the nature and types of maternal complications encountered in women with SLE who attempt pregnancy. In addition, although fetal outcome is reduced in these pregnancies, the incidence offailed pregnancies is much less than previously reported. Based on current insights it seems likely that
204
HAYSLETI
TABLE III. Anticardiolipin Antibody and Fetal Loss* Characteristic No. pregnancies No. pregnancies/patient % Fetal loss % Stillbirths % Anticardiolipin
Group 1 SLE & RFL a 36 4.6 69.0 12.0 83.0
Group 2 Group 3 Group 4 SLE RFL Normal 163 3.7 11.0 0.8 14.0
38 4.3 59.0 3.6 8.0
43 3.2 6.5 0 0
*Data from Deleze et al. 22 aRFL, recurrent fetal loss.
maternal complications are reduced and successful pregnancy outcome is enhanced by delaying pregnancy until SLE is inactive. Even when the activity ofSLE is quiescent before the onset of pregnancy, however, maternal complications are reported in 30%-50% of cases, and obstetrical complications involving preterm deliveries and small-for-gestational-age infants remain high. It seems likely that the improved outcome of pregnancy reported in recent case series is related to advances in obstetrical care, neonatal management, and, perhaps, advances in treatment ofSLE. Further improvements will depend on the introduction of a more specific and less hazardous treatment of the underlying disease. REFERENCES 1. Stern R, Fishman J, Brusman H, Kunkel HG. Systemic lupus erythematosus associated with Klinefelter's syndrome. Arthritis Rheum. 1977; 20:18-22. 2. Melez KA, Reeves JP, Steinberg AD. Regulation of the expression of autoimmunity in NZB x NZW Fl mice by sex hormones. J Immunopharmol.1978; 79:27-42. 3. Garsenstein M, Pollak VE, Kark RM. Systemic lupus erythematosus and pregnancy. N Engl J Med. 1962; 267:165-169. 4. Zulman JI, Talal N, Hoffman GS, Epstein WV. Problems associated with the management of pregnancies in patients with systemic lupus erythematosus. J Rheumatol. 1980; 7:37-49. 5. Lockshin MD, Reinitz E, Druzin ML, Murrman M, Estes D. Lupus pregnancy: Case-controlprospective study demonstrating absence of lupus exacerbation during or after pregnancy. Am J Med. 1984; 77:893-898. 6. Mintz G, Gutierrez G, Garcia-Alonso A, Karchmer S. Prospective study of pregnancy in patients with systemic lupus erythematosus. Arch Inter Med. 1991; 151:269-273.
7. Estes D, Larson DL. Systemic lupus erythematosus and pregnancy. Clin Obstet Gynecol. 1965; 8:307-321. 8. Hayslett JP, Lynn RI. Effect of pregnancy in patients with lupus nephropathy. Kidney Int. 1980; 18:207-220. 9. Bobrie G, Liote F, Houillier P, Grunfeld JP, Jungers P. Pregnancy in lupus nephritis and related disorders. Am J Kidney Dis. 1987; 9:339-343. 10. Houser MT, Fish AJ, Tagatz GE, Williams pp, Michael AF. Pregnancy and systemic lupus erythematosus. Am J Obstet Gynecol. 1980; 138:409-413. 11. Tozman ECS, Urowitz MB, Gladman DD. Systemic lupus erythematosus and pregnancy. J Rheumatol. 1980; 7:624-632. 12. Lockshin MD. Pregnancy does not cause systemic lupus erythematosus to worsen. Arthritis Rheum. 1989; 63:665-670. 13. Imbasciati E, Surian N, Bottino S, Cosci P, Colussi G, Ambroso GC, Lupus nephropathy and pregnancy: A study of 26 pregnancies in patients with lupus erythematosus and nephritis. Nephron. 1984; 36:46-51. 14. Hayslett JP. Effect of pregnancy in patients with SLE. Am J Kidney Dis. 1982; 2 (Suppll):223-228. 15. Wilcox AJ, Weinberg CR, O'Connor JF, Baird DD, Schlatterer JP, Canfield RE, Incidence of early loss of pregnancy. N Engl J Med. 1988; 319:189-194. 16. Varner MW, Meehan RT, Syrop CH, Strottmann P, Goplerud CPO Pregnancy in patients with sytemic lupus erythematosus. Am J Obstet Gynecol. 1983; 145:1025-1040. 17. Wong K-L, Chan F-Y, Lee CPo Outcome of pregnancy in patients with sytemic lupus erythematosus. Arch Intern Med. 1991; 151: 269-273. 18. Love PE, Santoro SA. Antiphospholipid antibodies: Anticardiolipin and the lupus anticoagulant in systemic lupus erythematosus (SLE) and in non-SLE disorders. Ann Intern Med. 1990; 1122:682-698. 19. Hughes GRV, Harris NN, Gharari AE. The anticardiolipin syndrome. J Rheumatol. 1986; 13:486-489. 20. Alarcon-Segovia D, Deleze M, Oria BS. Anticardiolipin antibodies and the anticardiolipin syndrome in systemic lupus erythematosus. Medicine. 1989; 68:353-365. 21. Silver SK, Adler L, Hickman A, Hageman JR. Anticardiolipin antibody-positive serum enhances endothelial cell platelet activating factor production. Am J Obstet Gynecol. 1991; 165: 1748-1752. 22. Deleze M, Alarc6n-Segovia D, Valdes-Macho E, Oria CV, Ponce deLeon S. Relationship between antiphospholipid antibodies and recurrent fetal loss in patients with systemic lupus erythematosus and apparently healthy women. J Rheumatol. 1989; 16: 768-772. 23. Loizou S, Byron MR, Englert HJ. Associations of quantitative anticardiolipin antibody levels with fetal loss and time of loss in systemic lupus erythematosus. Quart J Med. 1988; 68:525-531. 24. Kincaid-Smith P, Fairley KF, Kloss M. Lupus anticoagulant associated with renal thrombotic microangiopathy and pregnancy related renal failure. Quart J Med. 1988; 258:795-815.
The Effect of Systemic Lupus Erythematosus on Pregnancy and Pregnancy Outcome JOHN P. HAYSLETT Department ofMedicine, Section ofNephrology, Yale University School of Medicine, New Haven, Connecticut ABSTRACT: This review provides an analysis of reports published since 1980 on the effect of systemic lupus erythematosus (SLE) on pregnancy and pregnancy outcome. The question whether pregnancy increases clinical flares and the severity of flares in patients with SLE during pregnancy has not been resolved because of difficulty in defining exacerbations of SLE and of preeclampsia. An analysis of major detailed reports indicates that maternal complications are reduced in patients who are in clinical remission prior to the onset of pregnancy compared with women with persistent disease activity. Complications are observed in 30%-50% of patients with inactive disease at onset of gestation. After exclusion of spontaneous abortions during the first trimester, fetal survival was 85%-90% in most reported case series. The best outcomes were reported in patients with inactive disease at onset of pregnancy. It seems likely that some maternal complications and fetal wastage in this population are related to anticardiolipin antibodies. (Am J Reprod Immunol. 1992; 28:199-204.) Key words: Anticardiolipin antibodies, fetal outcome, maternal complications, pregnancy-induced exacerbations. INTRODUCTION
Systemic lupus erythematosus (SLE) is a disease that primarily affects women of child-bearing age. Whether SLE is affected by pregnancy and whether the likelihood for a successful outcome is reduced in patients with SLE are issues of concern. The scientific basis for answering these questions has improved recently because of numerous reports on relatively large series of patients with SLE who have experienced pregnancy. This review analyzes the larger series published since 1980, especially those that include detailed information. It should be recalled that the clinical literature related to SLE and pregnancy is confounded by wide differences in patient populations among individual reports. Some reports include cases with severe forms ofSLE, while others involve women with milder disease. In addition, there are series that include patients selected from multiple centers, as well as consecutive series of cases observed in a single clinic. In most reports, the quality of obstetrical care and neonatal management is unknown, and treatment regimens for SLE are not described. Nevertheless, these reports provide the database that is available for analysis.
Submitted for publication September 1, 1992; accepted September 8, 1992. Address reprint requests to John P. Hayslett, M.D., Department of Medicine, Yale School of Medicine, 333 Cedar Street - LMP 2073, New Haven, CT 06510-8056.
EXACERBATION OF SLE DURING PREGNANCY
The question whether pregnancy adversely affects SLE and causes a heightened incidence of clinical flares and/or increased severity of flares remains undetermined. On a theoretical basis an adverse effect on the underlying disease diathesis would be expected. There is evidence that estrogens have a potent influence on the immune system and on the initiation of SLE in man and in animal models ofSLE. The female:male ratio in man is 9:1 and SLE is reported in men with Klinefelter's syndrome. 1 Moreover, in the NZB x NZW F1 murine model ofSLE, the disease occurs predominantly in females but can be produced in males by administration of estrogens. 2 Sex hormones modulate the expression of autoimmunity and affect the concentration of antibodies to nucleic acid, the severity of glomerulonephritis, and, ultimately, mortality. In two retrospective clinical studes3.4 a marked increase in clinical exacerbations was observed during pregnancy and the puerperium, compared with the apparent incidence in the same patients prior to pregnancy. In a;rospective analysis, however, Lockshin and associates compared 21 patients with mild SLE who attempted to complete pregnancy with nonpregnant women with SLE matched for age, race, and severity of disease. The criteria used for clinical flares excluded new onset of proteinuria (four cases), regarded as probable preeclampsia, and new onset thrombocytopenia (eight cases in the pregnancy group compared with one in control patients). These investigators concluded that there was no difference in the incidence of exacerbations or in disease activity between pregnant and nonpregnant patients. Ifthese exclusive criteria for defining flares ofSLE had not been used in analysis, a marked increase in disease activity during pregnancy would have been recorded. A subsequent report by Mintz and associates" compared disease activity in a larger patient series of 102 pregnancies in 75 patients with a concurrent series of patients who did not experience pregnancy, and also found no difference in the apparent incidence of exacerbation. Details of these populations were not provided in their report. An explanation of the different results in these studies is confounded by the difficulty in establishing criteria for clinical flares of SLE, the absence of a firm clinical marker for preeclampsia, and the issue whether maternal complications thought to be caused by anticardiolipin antibodies should be regarded as flares of SLE. An attempt to resolve this question will be an important challenge for future studies. Regardless of the outcome, the reported incidence of exacerbations of 10%-75% during pregnancy and postpartum and the high fetal wastage of up to 40% in some reports highlight the concern for the potential of this group of patients to bear children successfully. In the series reported by Lockshin" with mild disease there was a 44% incidence of preterm deliveries and a fetal survival of only 68%.
200
HAYSLETT
ESTABLISHED SLE - INCIDENCE OF EXACERBATION REMISSION AT ONSET
ACTIVITY AT ONSET
100
100
80
80
60
60
40
40
20
20
o
o
o
Hayslett, Kid Inti 18:207, 1980
ELZI
Bobrie, Am J Kid Dis 9:339, 1987 ILl Houser, Am J OB/GYN 138:409, 1980 Fig.1.
e
Tozman, J Rheum 7:624, 1980
§
Lockshin, Arth & Rheum 32:665, 1989
Correlations of relapse or exacerbations ofSLE in relationship to activity of disease at onset of pregnancy.
MATERNAL COMPLICATIONS DURING PREGNANCY AND POSTPARTUM
In 1965 Estes and Larson 7 suggested, in an analysis of79 pregnancies after onset of disease in 36 women, that the effect of pregnancy on SLE correlated directly with the activity of disease at the time of conception. It is of interest that in this report, which spanned 20 years, glucocorticoid treatment was not available in more than onehalf of patients, and therefore the series may have included women with milder forms of SLE. In 35 pregnancies in which SLE was in remission at the start of pregnancy, relapse occurred in 26%, while remission was sustained in 74%. In contrast, SLE remained active in 37% and exacerbation occurred in 33% of pregnancies associated with active disease at conception. This correlation is supported by reports since 1980. Results of five studies are shown in Fig. 1 in which SLE predated pregnancy and in which sufficient data were available to assess clinical activity of SLE at onset of pregnancy. Hayslett and Lynn" reported on 65 selected pregnancies by 47 patients. These cases were selected by nephrologists and were managed in multiple centers in North America. In 80% of cases clinical renal disease was present before the index pregnancy. Following complete clinical remission ofSLE for at least 6 months prior to conception there was sustained remission in 65%, while 32% were complicated by relapse during pregnancy. Complete or partial remission occurred in all cases within 3 months of delivery. In contrast, the course of pregnancy associated with disease activity at conception was marked
by persistent or increased activity of SLE in 52%. Exacerbations were more severe in this group and clinical remission did not occur after delivery in four cases; one patient died within 6 months of delivery. Bobrie and associates" evaluated 53 pregnancies in 35 women after the onset of SLE. These cases were also a selected series and were initially managed in multiple medical centers in an urban center in Europe. Exacerbations were observed in 62% of women with signs of active disease at onset of pregnancy, compared with 34% in cases with remission at onset of pregnancy. In the former "group clinical flares were more severe, and five patients progressed rapidly to end-stage renal disease, all of whom had severe extrarenal manifestations. Houser and associatea'? described the clinical course of 18 pregnancies in 11 individuals with onset ofSLE in childhood. All cases were managed in a single medical center. Of 10 pregnancies associated with complete remission at conception, gestation was uncomplicated in 80%, while in the remaining eight pregnancies associated with clinical activity, increased clinical manifestations or worsening of renal disease occurred in 50%; none, however, exhibited rapidly progressive renal failure during follow-ur, which averaged 4 years. Tozman and associates 1 reported the course of 24 pregnancies in 18 women followed by one rheumatology clinic. Of 11 instances with inactive disease at onset of pregnancy only one exhibited relapse. In contrast, the disease remained active throughout pregnancy in 85% of pregnancies marked by clinical activity at onset. Lockshin'"
PREGNANCY AND SLE
described the course of 80 patients with SLE. Although clinical activity at conception was not reported, I assume, for the purpose of analysis, that 53 patients, or 66%, had inactive disease as they were not treated with corticosteroid, while 27 patients on corticosteroid had active ~is ease. As shown in Fig. 1, the recurrence or exacerbation rates were 37% in cases with activity at onset of pregnancy and 19% in those with inactive disease. In this computation one patient with onset of disease during pregnancy and two patients with recurrence of rash ~r termination of treatment were excluded from analysis. The report of Mintz and associates" was not included in the composite analysis as only 10% of patients had active disease at conception, and in most of these severity appeared to be mild. This large series with inactive SLE at onset of pregnancy, managed by one medical center, provides important insights into the incidence and types of recurrences of lupus activity in patients with inactive disease at the onset of pregnancy. Of the 92 pregnancies that began while in remission, 55 (60%) had an exacerbation during pregnancy, postpartum, or postabortion. In most cases the manifestations of SLE were mild and involved the cutaneous or musculoskeletal systems. Hematological manifestations of leukopenia and thrombocytopenia occurred in 35%, exacerbation or new onset of nephropathy was observed in 16%, and 15% exhibited central nervous system (CNS) disease. There were no maternal deaths. When the onset of SLE occurs during pregnancy or postpartum, clinical manifestations tend to be severe. Of nine patients in the report by Hayslett and Lynn," five had evidence of nephrotic syndrome, with renal insufficiency in two, and all had severe nonrenal manifestations. Similar hectic clinical courses in pregnancies associated with initial onset of SLE were described by Bobrie and associates" and Imbasciati and associates.l" Taken together these six publications describe SLE activity during pregnancy in over 300 gestations. Although caution should be exercised in extrapolating from these reports, as they represent heterogeneous populations of patients, these results suggest, as reported by Estes and Larson7 30 years ago, that the frequency and severity of recurrences or exacerbations during pregnancy tend to be more common and more severe when pregnancy commences in the setting of active SLE. In addition, they indicate that even in patients with inactive disease there is a high incidence of recurrent disease, including manifestations that involve the kidney and CNS.
201
FETAL OUTCOME
Inspection of published reports prior to 1980 indicates a live birth rate by women with established SLE of approximately 60%, after exclusion of therapeutic abortions.l" In most of these reports, as well as in recent publications, all spontaneous abortions were listed. Inclusion of spontaneous abortions during the first trimester inflates the incidence offetal death in SLE-related pregnancies because the spontaneous rate in normal women is over 20% in the first trimester. 15 Table I lists data on fetal outcome in 302 pre§nancies from nine reports published since 1980. 4,6,8-1 .13, 6,17 Publications from the same era that failed to provide sufficient detail on fetal outcome were omitted from analysis. The number of pregnancies listed for each report reflects only pregnancies in women with established SLE before the index pregnancies and in which the fetus survived beyond 12 weeks. The incidence of clinically evident lupus nephropathy, which was 60% or more in seven reports, is listed for each series because its presence correlates with the severity of SLE and because the adverse effects of renal disease and SLE on fetal outcome are probably additive. The incidence oflate abortions, stillbirths, and neonatal deaths have been combined for convenience. Inspection of these data indicates a wide range of preterm deliveries, defined as delivery before 37 weeks' gestation. In general, preterm deliveries were more common in patients with more active SLE, but there is no explanation for the wide variation between different studies, except for the possibility of different attitudes toward delivery by Caesarian section or induction oflabor among different medical centers. In eight ofthe nine reports fetal survival exceeded 80%, despite the high incidence of lupus nephritis and often hectic maternal courses described in most of these studies. The reason for the low survival rate of 56% in the Imbasciati et al. 13 series is not clear, but may relate to referred patients with more severe systemic disease; severe renal insufficiency was reported in only one patient. Not suprisingly, when fetal outcome was correlated with maternal SLE activity, survival rate was higher in patients with a lower incidence of recurrence or exacerbation, as shown in Fig. 2. These data were abstracted from four reports8-11 of patients with established SLE prior to conception, where sufficient detail was provided to make such a correlation. In the group with inactive
TABLE I. Pregnancy Outcome in Patients With Established SLE*
Author
Hayslett" Bobrie? Tozman!' Houser!" Zulman" Mintz" Varner 16 Imbasciati 13 Wong l ?
No. pregnancies
% Renal
% Preterm
disease
births
% Deaths
% Fetal andneontal survival
47 38 18 14 24 102 26 16 17
80 100 61 100 67 59 21 100 12
4 13 8 21
13 13 11 7 4 11 0 38 0
81 87 89 93 83 89 100 56 100
"Deaths include second and third trimester abortions, stillbirths, and neonetal deaths.
49 12 69 41
202
HAYSLETT
ESTABLISHED SLE - FETAL SURVIVAL REMISSION AT ONSET
ACTIVITY AT ONSET
100 80
'E 60 Q) ~
Q)
a.. 40 20
o
Hayslett, Kidney Inter 18:207,1980
~ Dobrie, Am J Kid Dis 9:339, 1987
. . Houser, Am J Obstet Gynec 138: 409, 1980 ~ Tozman, J Rheum 7:624, 1980
Fig. 2. Fetal survival in patients with SLE in relationship to activity of disease at onset of pregnancy.
disease at conception, and therefore fewer flares during pregnancy, the live birth rate was 88%-100%. In contrast, in women with active disease at onset of gestation the live birth ranged between 50% and 82% in the same series. Maternal disease activity is usually reported to be severe when SLE initially began during pregnancy or immediately postpartum. In the three reports 8,9,16 that provided information on this subgroup, fetal survival was reduced to 50%-65%, as shown in Fig. 3. The effect of maternal SLE on fetal growth and development is not characterized solely by survival rate. The data in Table II were derived from the large series of Mintz and associates" managed by one center with modern perinatal care facilities. They demonstrate high rates of preterm deliveries and intrauterine growth impairment even in women with predominantly inactive disease at onset of pregnancy. Not surprisingly, the incidence of small-forgestational-age infants was higher in preterm deliveries and averaged 23% in the series as a whole. ROLE OF ANTICARDIOLIPIN ANTIBODIES IN MATERNAL COMPLICATIONS AND FETAL WASTAGE
Recent studies have suggested that antiphospholipid antibodies, including anticardiolipin and the lupus anticoagulant, play an important role in the pathogenesis
of vascular lesions in patients with SLE. 18,19 The average prevalence of anticardiolipin antibody is reported to be 44% in SLE, and that of the lupus anticoagulant, 34%.18 There is strong evidence that high antibody titers correlate with arterial and venous thrombosis, hemolytic anemia, thrombocytopenia, and recurrent fetal 10ss.20 Both types of antibodies recognize similar antigenic determinants, and are postulated to enhance platelet clearance by binding to membrane phospholipids or by promoting platelet activation.l" Recent evidence indicates that anticardiolipin enhances platelet-activating factor by vascular endothelial cells. 21 These types of mechanisms are probably responsible for the formation of platelet thrombi on vessel walls. In addition, activated platelets have a critical procoagulant role and serve as a nidus for membrane-based coagulation reactions involving the generation of thrombin and the formation of fibrin. It is thought that spontaneous abortion is caused by venous thrombosis in the placenta. The data, shown in Table III were derived from the report by Deleze and associates.~2In Group 1, comprised of patients with SLE and recurrent loss of three or more products of conception, nearly 80% of pregnancies were unsuccessful compared with patients with SLE who did not have a history of recurrent unsuccessful pregnancies (Group 2). The incidence of anticardiolipin was 83% in
PREGNANCY AND SLE
203
ONSET OF SLE DURING PREGNANCY OR POST PARTUM - FETAL SURVIVAL 100 80
C (l)
60
~
(l)
o,
40
20
o
Hayslett, Kidney Inter 18:207,1980
~ Dobrie, Am J Kid Dis 9:339, 1987
G Fig. 3.
Varner, Am J Obstet Gynec 145:1025, 1983
Fetal survival in patients with SLE who experienced the onset of disease during pregnancy or postpartum,
Group 1 and 14% in Group 2. The groups with SLE are contrasted with healthy patients with (Group 3) and without (Group 4) recurrent fetal wastage. In this study the likelihood of fetal loss in patients with SLE and anticardiolipin antibodies was approximately 10 times greater than in patients without antibody titers. In nearly all cases with fetal wastage the antibody was the IgG isoform. In another study of84 patients without a ~revi ous history of repeated spontaneous abortions," IgG anticardiolipin antibodies correlated with a spontaneous abortion rate of 59% in 46 patients, compared with the rate of 25% in 38 patients without the antibody. In addition, abortion occurred later in gestation (17.4 ± 7.1 wk) than in patients without antibodies (12.5 ± 7.1 wk). There is no evidence that treatment with glucocorticoids or antiplatelet agents, such as aspirin, reduces the incidence of fetal loss due to anticardiolipin antibodies. Because high titers of anticardiolipin antibodies have been associated with vascular and hematological complications in nonpregnant patients with SLE, it seems likely that some maternal complications during pregnancy are also caused by this mechanism. In the series reported by Deleze and associates.P there were statistically higher rates of thrombocytopenia, venous throm-
bosis, and convulsions in patients with anticardiolipin antibodies and recurrent fetal loss compared with cases without antibodies. In addition, recent studies suggest that antibody-related thrombosis may also playa role in exacerbations of renal disease during pregnancy. KincaidSmith and associates'" described the course of pregnancy in 12 women with circulating antiphospholipid antibodies, four of whom had SLE. Renal biopsy showed fibrin thrombi in glomeruli and arteries. Renal function was reduced in six of these patients and only two of23 pregnancies were successful. Although it seems clear that anticardiolipin antibodies play an important role in the mechanism offetal wastage and maternal complications in SLE, the relative importance of this factor has not been defined. There is, for example, a wide discrepancy between the rate offetal loss predicted from the reported prevalence of anticardiolipin antibodies in SLE of about 40% and an associated spontaneous abortion rate of 50%-60%, and the observed incidence of late abortions and stillbirths in series ofpregnant women with SLE oflO%-15%, reported before discovery of anticardiolipin antibodies. Clearly future studies are needed that include well characterized unselected patients; serial measurements of anticardiolipin before and during gestation; and controls matched for age, disease, and clinical history.
TABLE II. The Effect of SLE on Fetal Outcome* Characteristic
ActiveSLE
Inactive SLE
No. pregnancies
51
% Preterm deliveries % Abortions % Stillbirths % Small for gestational age
59
51 39
14
20
% Term deliveries
27
'Data from Mintz et al. 6
3
2
23
41
CONCLUSION
In summary, important strides have been made during the past decade in understanding the nature and types of maternal complications encountered in women with SLE who attempt pregnancy. In addition, although fetal outcome is reduced in these pregnancies, the incidence offailed pregnancies is much less than previously reported. Based on current insights it seems likely that
204
HAYSLETI
TABLE III. Anticardiolipin Antibody and Fetal Loss* Characteristic No. pregnancies No. pregnancies/patient % Fetal loss % Stillbirths % Anticardiolipin
Group 1 SLE & RFL a 36 4.6 69.0 12.0 83.0
Group 2 Group 3 Group 4 SLE RFL Normal 163 3.7 11.0 0.8 14.0
38 4.3 59.0 3.6 8.0
43 3.2 6.5 0 0
*Data from Deleze et al. 22 aRFL, recurrent fetal loss.
maternal complications are reduced and successful pregnancy outcome is enhanced by delaying pregnancy until SLE is inactive. Even when the activity ofSLE is quiescent before the onset of pregnancy, however, maternal complications are reported in 30%-50% of cases, and obstetrical complications involving preterm deliveries and small-for-gestational-age infants remain high. It seems likely that the improved outcome of pregnancy reported in recent case series is related to advances in obstetrical care, neonatal management, and, perhaps, advances in treatment ofSLE. Further improvements will depend on the introduction of a more specific and less hazardous treatment of the underlying disease. REFERENCES 1. Stern R, Fishman J, Brusman H, Kunkel HG. Systemic lupus erythematosus associated with Klinefelter's syndrome. Arthritis Rheum. 1977; 20:18-22. 2. Melez KA, Reeves JP, Steinberg AD. Regulation of the expression of autoimmunity in NZB x NZW Fl mice by sex hormones. J Immunopharmol.1978; 79:27-42. 3. Garsenstein M, Pollak VE, Kark RM. Systemic lupus erythematosus and pregnancy. N Engl J Med. 1962; 267:165-169. 4. Zulman JI, Talal N, Hoffman GS, Epstein WV. Problems associated with the management of pregnancies in patients with systemic lupus erythematosus. J Rheumatol. 1980; 7:37-49. 5. Lockshin MD, Reinitz E, Druzin ML, Murrman M, Estes D. Lupus pregnancy: Case-controlprospective study demonstrating absence of lupus exacerbation during or after pregnancy. Am J Med. 1984; 77:893-898. 6. Mintz G, Gutierrez G, Garcia-Alonso A, Karchmer S. Prospective study of pregnancy in patients with systemic lupus erythematosus. Arch Inter Med. 1991; 151:269-273.
7. Estes D, Larson DL. Systemic lupus erythematosus and pregnancy. Clin Obstet Gynecol. 1965; 8:307-321. 8. Hayslett JP, Lynn RI. Effect of pregnancy in patients with lupus nephropathy. Kidney Int. 1980; 18:207-220. 9. Bobrie G, Liote F, Houillier P, Grunfeld JP, Jungers P. Pregnancy in lupus nephritis and related disorders. Am J Kidney Dis. 1987; 9:339-343. 10. Houser MT, Fish AJ, Tagatz GE, Williams pp, Michael AF. Pregnancy and systemic lupus erythematosus. Am J Obstet Gynecol. 1980; 138:409-413. 11. Tozman ECS, Urowitz MB, Gladman DD. Systemic lupus erythematosus and pregnancy. J Rheumatol. 1980; 7:624-632. 12. Lockshin MD. Pregnancy does not cause systemic lupus erythematosus to worsen. Arthritis Rheum. 1989; 63:665-670. 13. Imbasciati E, Surian N, Bottino S, Cosci P, Colussi G, Ambroso GC, Lupus nephropathy and pregnancy: A study of 26 pregnancies in patients with lupus erythematosus and nephritis. Nephron. 1984; 36:46-51. 14. Hayslett JP. Effect of pregnancy in patients with SLE. Am J Kidney Dis. 1982; 2 (Suppll):223-228. 15. Wilcox AJ, Weinberg CR, O'Connor JF, Baird DD, Schlatterer JP, Canfield RE, Incidence of early loss of pregnancy. N Engl J Med. 1988; 319:189-194. 16. Varner MW, Meehan RT, Syrop CH, Strottmann P, Goplerud CPO Pregnancy in patients with sytemic lupus erythematosus. Am J Obstet Gynecol. 1983; 145:1025-1040. 17. Wong K-L, Chan F-Y, Lee CPo Outcome of pregnancy in patients with sytemic lupus erythematosus. Arch Intern Med. 1991; 151: 269-273. 18. Love PE, Santoro SA. Antiphospholipid antibodies: Anticardiolipin and the lupus anticoagulant in systemic lupus erythematosus (SLE) and in non-SLE disorders. Ann Intern Med. 1990; 1122:682-698. 19. Hughes GRV, Harris NN, Gharari AE. The anticardiolipin syndrome. J Rheumatol. 1986; 13:486-489. 20. Alarcon-Segovia D, Deleze M, Oria BS. Anticardiolipin antibodies and the anticardiolipin syndrome in systemic lupus erythematosus. Medicine. 1989; 68:353-365. 21. Silver SK, Adler L, Hickman A, Hageman JR. Anticardiolipin antibody-positive serum enhances endothelial cell platelet activating factor production. Am J Obstet Gynecol. 1991; 165: 1748-1752. 22. Deleze M, Alarc6n-Segovia D, Valdes-Macho E, Oria CV, Ponce deLeon S. Relationship between antiphospholipid antibodies and recurrent fetal loss in patients with systemic lupus erythematosus and apparently healthy women. J Rheumatol. 1989; 16: 768-772. 23. Loizou S, Byron MR, Englert HJ. Associations of quantitative anticardiolipin antibody levels with fetal loss and time of loss in systemic lupus erythematosus. Quart J Med. 1988; 68:525-531. 24. Kincaid-Smith P, Fairley KF, Kloss M. Lupus anticoagulant associated with renal thrombotic microangiopathy and pregnancy related renal failure. Quart J Med. 1988; 258:795-815.
