Female patients with systemic lupus erythematosus (SLE)-associated antiphospholipid syndrome (APS) have a lower ovarian reserve than either primary APS or SLE patients

Introduction

Antiphospholipid syndrome (APS) is characterized by the presence of antiphospholipid (aPL) antibody, is an autoimmune disease characterized by thrombosis, pregnancy complications [1], and is significantly associated with systemic lupus erythematosus (SLE). SLE is an autoimmune disease that involves an attack by the immune system on its own tissues, causing widespread inflammation and tissue damage in the affected organs. APS can occur as a primary condition or in the setting of SLE or another systemic autoimmune disease, and APS and SLE are closely related autoimmune diseases with overlapping clinical and biological characteristics [2]. Cervera et al. conducted a 10-year follow-up study of 1,000 patients with APS from 13 European countries and reported that 36.2 % of APS patients possessed associated SLE [3].

Secondary APS (SAPS) is observed in the context of a coexisting connective tissue disease such as SLE or rheumatoid arthritis (RA), and a majority of SAPS cases occur in conjunction with SLE. The 2006 International Consensus Statement on updating the classification criteria for APS did not recommend the use of the term SAPS. The statement instead recommended the documentation of the presence or absence of SLE which is more advantageous for disease classification rather than categorizing patients into primary and secondary APS [4]. Many experts now prefer the term SLE-APS, and SLE-APS patients are more likely to manifest thrombocytopenia than those with primary APS [5].

Although autoimmunity correlates with pregnancy complications [6], the use of therapeutic immunomodulators makes pregnancy relatively safe [7]. APS is more common in young women of reproductive age [8], and while the relationship between APS and female fertility is questioned [6, 7], adverse pregnancy outcomes constitute a major clinical manifestation of APS. Female patients of reproductive age with APS show a high risk of recurrent early pregnancy loss and late fetal loss [9]. Yamakami et al. [10] reported lower antral follicle count (AFC) and anti-Müllerian hormone (AMH) levels in individuals with primary APS, suggesting reduced ovarian reserve (OR). Franco et al. [11] demonstrated that of 35 patients with SLE-APS, 51.4 % revealed a prior thromboembolic event, 40 % had a previous pregnancy loss and 8.5 % exhibited both. Ünlü et al. [12] reported that compared to SLE patients without aPL, those with aPL showed a higher prevalence of thrombosis and pregnancy morbidity. There may, however, be some differences in ovarian function among women with PAPS, SLE-APS, and SLE.

We herein determined OR using traditional indicators such as AFC, follicle-stimulating hormone (FSH) and AMH [13, 14]. Both AMH and INHB are members of the transforming growth factor (TGF)-β family that are secreted by ovarian granulosa cells [15, 16]. AMH mediates follicular development via an FSH-independent mechanism [10], while INHB suppresses FSH release [17]. Serum AMH also reflects a strong positive correlation with AFC in the general population [18, 19]. We compared differences in ovarian function in women with PAPS, SLE-APS and SLE, which could potentially explain pregnancy complications in these women.

Materials and methods

Results

ROC curves of AMH to predict ovarian reserve decline of APS patients

The ROC curve reflects the relationship between sensitivity and specificity. Our analysis of the ROC curve indicated that AMH level was highly accurate in predicting ovarian reserve decline (AUC=0.935; 95 % CI, 0.893–0.978; p<0.001; Figure 1). Based on the ROC curve, we identified an AMH value of 0.67 ng/mL as the threshold for predicting ovarian reserve decline in APS patients.

Figure 1:

ROC curves of AMH to predict ovarian reserve decline. AUC of AMH of APS patients in ovarian reserve decline is 0.935 (95 % CI: 0.893–0.978; p<0.001).

AMH, INHB, AFC, E2, and P differ among PAPS, SLE-APS, and SLE patients

We compared the immune-related blood tests and total (Figure 2) AFCs among HS, PAPS, SLE-APS, and SLE groups (Table 3), and demonstrated that AMH and AFC levels were lower in APS, SLE-APS, and SLE patients compared with HS, and that FSH levels in the disease groups were higher than in the HS group, but that E2 was reduced.

Figure 2:

Correlation analysis of AMH with INHB, aCL antibody, age, and AFC in APS patients.

Table 3:

Comparison of blood indices and antral follicle count of HS, PAPS, SLE-APS, and SLE groups before treatment.

	HS (n=48)	PAPS (n=37)	SLE-APS (n=69)	SLE (n=52)	p-Value
Age, years	28 (26–33)	28 (26.5–30)	30 (28–31)	29 (26–31.75)	0.274
AMHa, ng/mL	2.61 (1.50–3.35)	1.69 (1.06–2.02)	0.45 (0.13–0.78)	0.89 (0.33–1.32)	0.000
INHBb, pg/mL	56.55 (34.28–66.53)	36.6 (32.8–45.2)	20.2 (15.3–29.5)	36.75 (22.08–56.23)	0.000
AFCb, N	15 (10–17)	11 (10–13)	7 (5–10)	9.5 (7–12)	0.000
FSHb, IU/L	6.02 (4.44–8.18)	7.88 (5.66–8.94)	9.97 (8.44–11.3)	8.35 (5.46–9.87)	0.000
E2b, pmol/L	182 (117–321.3)	95 (76.5–115)	56 (34–87.5)	87 (65.5–121.8)	0.000
Pc, nmol/L	2 (1.64–2.22)	1.7 (1.4–1.99)	1.66 (1.36–1.88)	1.86 (1.48–2.0)	0.000
T, nmol/L	1.68 (1.47–1.99)	1.45 (1.22–1.89)	1.48 (1.26–1.86)	1.91 (1.45–2.35)	0.309
β2-GPI, AU/mL	5.35 (4.93–6.5)	5.8 (5–10.3)	6.15 (5.1–10.35)	5.45 (5–8.85)	0.261
aCLc, PLU/mL	6.35 (5–8.38)	12.2 (6.3–37.6)	20.1 (6.3–97.85)	7.5 (5–19.8)	0.000

1. Data are expressed as median (Q25–Q75); HS, healthy subjects; ^ap <0.05, the four groups were compared in pairs; ^bp <0.05, the four groups were compared in pairs except for PAPS and SLE groups; ^cp <0.05, PAPS and SLE-APS groups were compared with healthy controls.

Discussion

APS is an autoantibody-mediated disorder that primarily affects the platelets, resulting in recurrent thrombosis and its sequelae, including pregnancy loss. APS is principally diagnosed between 20 and 50 years of age and affects 3–5 times more women than men. APS is considered to be a systemic autoimmune disease and may coexist with other autoimmune diseases, with SLE being the most frequently reported. A decline in OR is frequently uncovered in patients with autoimmune diseases such as SLE [27]. A meta-analysis by Wahl et al. [28] reported that patients with SLE and APS were at an approximately six-fold increased risk of venous thromboembolism compared to women without APS. We now appreciate that AMH is a reliable marker of OR that can predict early ovarian follicle loss and menopausal onset in the general population [29]. For example, Yamakami et al. [10] reported reduced AFC and AMH levels in individuals with primary APS, suggesting an OR decline, and Luo et al. [30] also demonstrated that SLE was associated with low AMH levels. In our study, we classified the three diseases by clinical and serologic characteristics (Table 1), and determined that the OR decline in APS, SLE-APS, and SLE patients differed in terms of blood indices.

Serum AMH is a cost-effective and reliable marker of OR that is highly and positively correlated with AFC. In our study, we depicted serum AMH levels as positively correlated with AFC and INHB levels in APS patients (Figure 2). Therefore, serum INHB levels during the early follicular phase of the menstrual cycle directly reflect AFC. We also found that APS patients manifested an AMH level <0.5 ng/mL, which accounted for a high proportion of the total (40.17 %). The mean AMH level and AFC were lower in the PAPS, SLE-APS, and SLE groups relative to the HS group (p<0.05) (Table 3), congruent with the findings of Yamakami et al., who reported reduced AFC and AMH levels with primary APS [10]. Martins et al. [31] showed lower AMH levels in the majority of patients with SLE compared to HS. In our study, multivariate logistic regression analysis revealed that low AMH levels constituted independent risk factor for the decline in OR in APS patients (Table 2), and our ROC curve showed that AMH was highly accurate in predicting OR failure (Figure 1). We posit that AMH may be used to evaluate OR in APS patients, we identified an AMH value of 0.67 ng/mL as the appropriate threshold for predicting OR failure in APS patients.

Autoimmunity is one of the factors involved in premature ovarian failure (POF) [32]. Morel et al. [33] found that AMH levels are low in SLE patients and decrease significantly with age and cyclophosphamide (CYC) exposure. CYC is an alkylating agent which is widely used to treat SLE, using CYC often results in a lower cumulative dose, particularly with oral CYC, may ultimately reach cumulative doses associated with significant, long-term toxicity. Clowse et al. [34] think administration to a female patient of several pulse CYC regimens over the course of SLE poses a risk of premature ovarian insufficiency particularly for women in their 30 and 40 s, even if the risk from a single treatment is relatively small. Our patients did not receive CYC before diagnosed, and this therefore did not contribute to the reduced OR. POF primarily presents with a decline in OR, and is characterized by a reduction in AFC, AMH, and INHB levels, and in augmented FSH levels [35]. We observed similar results for APS, SLE-APS, and SLE patients compared to the HS group. In addition, E2 and P levels also decreased in each group (Table 3). FSH is required for the production and maturation of ovarian oocytes, and FSH levels drop on approximately day three of the menstrual cycle. FSH testing on day 3 may therefore be useful for evaluating OR, and in the present study we evaluated the FSH level on days 2–4. In addition to FSH, E2 levels should also be measured. E2 maintains the reproductive system, but high E2 levels suppress FSH. Furthermore, AMH decreases the activities of FSH-induced aromatase in granulosa cells [36]; in turn, high FSH levels is provoked by a reduction in AMH. In our study we found that among the disease groups, lower AMH and INHB levels correlated with higher FSH and low and steady E2 levels. The elevated FSH levels may be principally due to the lack of inhibition by AMH and INHB.

E2 plays a negative role in autoimmune diseases and SLE exhibits a significant female bias, particularly when E2 is at its peak [37]. In one study entailing a mouse model, the authors found that estrogen worsened the outcomes of autoimmune disease treatment [38]. However, at the beginning of menstruation (i.e., days 2–5), we discerned that the E2 level was significantly lower in the three disease groups than in the HS group, and SLE-APS manifested a particularly lower, while the E2 levels did not increase during this period and remained at low levels. In an article published in 2003, McMurray et al. found that the E2 levels in female SLE patients were different from those of their HS group; compared to the latter, the E2 level of SLE patients was either low, high, or showed no difference [39]. However, due to the lack of uniform collection strategies for blood testing in terms of the day of the menstrual cycle and detection methods used at the time of these disparate studies, there may be variations in these results. The difference in the collection times of blood samples for E2 measurements during the menstrual cycle appears to constitute the principal reason for the difference in E2 levels. In addition, the selection of assay methods and patient population may also be responsible for the observed differences. Further research is therefore required to determine the reasons for the particularly low E2 level in our SLE-APS patients on days 2–5 of the menstrual cycle.

P is the upstream precursor of T and E2, is necessary to maintain pregnancy, and a decline in P levels stimulates uterine contractions to initiate labor in humans, resulting in delivery [40]. In agreement with our study, Verthelyi et al. [41] found that the P concentration was not different in SLE patients compared to HS; however, we did show that the P level was reduced in patients with APS and SLE-APS. Velayuthaprabhu et al. [42] found that serum E2 and P levels were significantly reduced in studies in mice, suggesting that human β2-GPI may be involved in APS pregnancy failure through its effects on serum hormones. Therefore, the observed drop in P levels may be due to the presence of related antibodies in APS patients, leading to the decline in ovarian function.

APS is an autoimmune disease with detectable lupus anticoagulant (LA), aβ2GPI antibody or aCL antibody in serum, and approximately 20–40 % of SLE patients are aPL positive [12]. There is evidence showing that the presence of aPL is associated with attenuated AMH levels, while some studies show no relationship between them [43]. The AMH levels in our disease groups were lower than in the HS, with AMH reduced in the SLE group relative to the PAPS group, and lower in the SLE-APS group compared to the other two groups. The aCL level was higher in the disease groups relative to HS group: it was higher in the PAPS group compared to the SLE group, and the SLE-APS group had a higher aCL level than the other two groups (Table 3). While serum anti-B2-GPI antibody in APS patients may significantly improve the predictive rate and diagnostic value of thrombotic complications, it also possesses a high diagnostic value. However, the sensitivity of our B2-GPI antibody was not high and there was little difference in the levels among groups. Pekonen et al. [44] found circulating antibodies against ovarian tissue in the serum of lupus patients, suggesting the existence of autoimmune ovarian lesions in SLE patients [45]. Chen et al. [46] determined that antibodies were directed against pre-ovulatory follicle cells but not to corpus luteum-specific antigens, and this may impact OR in PAPS patients. We postulate that aPL antibodies reduce AMH, and that the reduction in AMH levels is more significant in cases of superimposed antibodies. Furthermore, reduced serum AMH levels may be due to insults against the ovarian microvasculature such as a thrombotic event. Likewise, the AFC of the SLE-APS group was the lowest among the groups, implying that growth of the primordial follicle cohort was impaired within an aPL milieu, and that the ovarian blood supply predisposed the primordial follicle cohorts to ischemic microvascular events.

There were some limitations to the present study. We collected as the total sample size all related cases from Jinhua Central Hospital over the past four years, and analyzed the blood indicators of patients who had been diagnosed but not yet treated with drugs. However, some SLE-APS patients were enrolled on the basis of their SLE disease, and these patients generally underwent drug treatment for SLE; therefore, several SLE-APS blood indicators likely contained some drug interference. The number of collected specimens was also limited and the changes in various indicators of the menstrual cycle were not monitored. Therefore, the dynamic changes in indices occurring over the menstrual cycle could not be compared. For example, the E2 level in the disease group (particularly in the SLE-APS group) was lower at approximately day 3, but the E2 levels for subsequent periods were not determined. Median AMH of the healthy subjects which is lower than typically observed in healthy women of the same age. The inconsistent data regarding AMH were also likely related to the differences in the phase of the menstrual cycle during which our testing was conducted.

Conclusions

In this study we ascertained that the AMH and AFC levels were lower in APS, SLE-APS, and SLE patients compared with HS patients, low AMH levels and high LH constituted independent risk factor for the decline in OR in APS patients, and that AMH and AFC levels were lower in SLE-APS patients relative to other groups. FSH and LH levels in the disease groups were also higher than in HS women, but the E2 level was reduced and exhibited no inhibitory effect on the gonadotropins, which was potentially caused by characteristically low AMH levels. The low levels of AMH may have been provoked by aCL or other related antibodies acting on the ovarian microvasculature, and this may have produced the lowest OR in SLE-APS patients.