Systematic review of the literature informing the systemic lupus erythematosus indicators project: Reproductive health care quality indicators

Authors


Abstract

Objective

Systemic lupus erythematosus (SLE) primarily affects women of reproductive age. Here we summarize the scientific evidence supporting recently developed quality indicators (QIs) pertaining to reproductive health.

Methods

We used a modification of the RAND/UCLA Appropriateness Method to develop QIs for SLE. We performed systematic reviews of the literature pertaining to each proposed indicator. Three indicators focusing on reproductive health were included in the final set. Relevant literature was presented to an expert panel, who rated the validity and feasibility of the indicators.

Results

Three QIs were rated as valid and feasible. These indicators specifically address laboratory testing during pregnancy in SLE, the treatment of antiphospholipid antibody syndrome, and counseling for drugs with teratogenic potential.

Conclusion

We used a rigorous method to develop reproductive health QIs for SLE. In the future, these indicators can be used in the assessment and delivery of care to patients with SLE.

INTRODUCTION

Systemic lupus erythematosus (SLE) is an autoimmune disease with protean manifestations characterized by periods of flare and remission. SLE is commonly diagnosed during the childbearing years; therefore, issues surrounding reproductive health are central in clinical practice.

In 1996, an initiative launched by the Institute of Medicine encouraged the assessment and improvement of health care quality, and efforts to improve the quality of patient care have extended to those with diagnosed rheumatic illness (1–3). In the US, the most commonly used tools to assess health care quality have taken the form of quality indicators (QIs), which are defined as “retrospectively measurable elements of practice performance for which there is evidence or consensus that can be used to assess the quality of care provided and hence change it” (4). Distinct from guidelines, QIs aim to represent a measurable standard of care across a specific patient population (3). In rheumatology, the development of QIs has provided insight into the management of gout, osteoarthritis, and rheumatoid arthritis using evidence-based literature searches and reproducible methodology (1–3, 5–7).

In order to assess the quality of care provided to patients with SLE, we recently developed a set of QIs employing the RAND/UCLA Appropriateness Method (8). Three of these indicators pertain to pregnancy and reproductive health. Here we present the results of the systematic literature reviews that served as the scientific foundation for developing the SLE reproductive health indicators.

MATERIALS AND METHODS

The methodology used to arrive at the full set of 20 SLE QIs is presented in detail elsewhere (8, 9) and only discussed briefly here. Below we describe the methods and results of our systematic literature reviews to derive the scientific evidence base for the reproductive health QIs.

We initially devised a preliminary list of potential QIs after a systematic review of the literature of guidelines and recommendations for SLE care. Each QI was structured in an IF-THEN-BECAUSE format, where the IF component defined the eligible population, the THEN component clearly defined the process of care to be performed by health care providers, and the BECAUSE component summarized the expected health benefits for the patient. Five QIs pertinent to pregnancy and reproductive health were proposed (see Supplementary Appendix A, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658).

As detailed in the article outlining the specifics of our formal group methods, the proposed indicators and literature review were then distributed to an advisory panel. After a face-to-face meeting and 2 rounds of electronic voting, the list of reproductive health indicators was reduced to 3. These QIs progressed to the next phase of the project, which utilized the RAND/UCLA Appropriateness Method (detailed in the methods article [8]) to derive a final QI set. Here we present the systematic literature review results presented to the final expert panel regarding the 3 reproductive health QIs.

Medline and EMBase were queried for abstracts from 1960 to 2008, with the help of professional librarians. The specific search strategies are available from the corresponding authors for review. For all of the searches, we excluded non-English studies, single case reports, review articles, and editorials. Titles and abstracts were reviewed for relevance independently by 2 authors, and disagreements were resolved through discussion. Full-text articles were then retrieved for further review. The results of the systematic literature review are discussed below.

RESULTS

QI 1.

IF a patient with SLE is pregnant, THEN SSA, SSB, and antiphospholipid antibodies (aPL) should be documented in the medical record, BECAUSE these assays may inform the patient and physician about the risks associated with pregnancy and influence the type and frequency of monitoring.

The scientific evidence pertaining to this QI was searched as 2 separate research questions: 1) Does screening for SSA and SSB antibodies in a pregnant woman with SLE change outcomes or provide information vital to patient care? and 2) Does screening for aPL in pregnant women with SLE change outcomes or provide information vital to patient care?

In part 1, we assessed the relevance of SSA and SSB antibody screening during pregnancy and focused on clinical outcomes in seropositive mothers. Our search yielded 105 articles from Ovid/Medline and 18 additional articles from EMBase. Of these 123 articles, 45 were selected for full-text review. Nine articles were deemed appropriate to the research question and used in the summary below. Evidence utilized in the summary included observational research only. No randomized clinical trials, systematic reviews, or meta-analyses were found.

Epidemiology and clinical syndromes associated with SSA and SSB antibodies.

Sjögren's antibodies (i.e., SSA and SSB) are common in women in SLE, with an estimated lifetime incidence of 67% and 49%, respectively (10). The transit of these antibodies passively through the placenta can induce a neonatal lupus syndrome (NLS). This syndrome has been associated with the development of congenital heart block (CHB) in the fetus and/or photosensitive rash and hematologic and hepatic abnormalities in the newborn (11). The incidence of CHB in the fetus of a seropositive woman has been estimated to be between 2% and 5% (12, 13), and the risk increases to approximately 16–25% when a seropositive mother has previously given birth to a child with CHB (14–16). The incidence of NLS in the offspring of seropositive mothers has not been clearly determined, but one study of 128 infants found either hematologic, dermatologic, or hepatic abnormalities in 27%, 16%, and 26%, respectively (17).

Monitoring and treatment of CHB.

CHB, which occurs as the fetal conduction system develops, is most often seen between 16 of 24 weeks of gestation (16, 18) and is associated with a 20% mortality rate by the end of infancy (16). Therefore, most guidelines recommend monitoring of fetal PR interval by echocardiogram weekly for seropositive mothers between 16 and 26 weeks of gestation and biweekly between 26 and 32 weeks of gestation (16, 19). Infants with severe CHB may require early delivery for the placement of a pacemaker (16).

Although third-degree heart block appears irreversible, a therapeutic window may exist in some cases for first- or second-degree heart block, and several cases have been reported showing improvement with the administration of steroids (Table 1). However, this literature is limited by the very small number of cases reported and the lack of a control group.

Table 1. Studies examining glucocorticoid therapy in congenital heart block (CHB)
Author, year (ref.)ObjectiveStudy designPopulationResults
Buyon et al, 1995 (61)To determine the feasibility and effectiveness of prenatal therapy for CHBMailed questionnaires and chart review in identified cases72 pregnancies with CHB19 pregnancies were treated with fluorinated steroids after identification of CHB
    1 with second-degree block reverted to sinus rhythm and 2 with third-degree block improved
Shinohara et al, 1999 (62)Study of 40 SSA-positive women who were followed for CHB and the effects of treatment with steroidsProspective cohort and retrospective case series87 offspring of 40 SSA-positive women were followed0 of 26 infants with SSA-positive mothers who were treated with steroids before 16 weeks of gestation developed CHB
   Only 2 patients with SLE15 of 61 infants without therapy developed CHB
   Many patients were prophylactically treated with steroids when at high risk of CHBOnce CHB was diagnosed (usually at 20 weeks of gestation), steroid therapy did not reverse the pathology
Saleeb et al, 1999 (63)To compare intervention with fluorinated glucocorticoids with the natural history of untreated CHBReview of a national registry for neonatal lupusThis study compared 18 pregnancies in a fluorinated glucocorticoid-treated group with 22 pregnancies in an untreated group with CHBFluorinated steroids do not seem to reverse third-degree heart block
    3 fetuses in the treated group and 2 fetuses in the untreated group with second-degree heart block progressed to third-degree block
    4 fetuses with treated second-degree heart block reverted to first-degree block, perhaps indicating some benefit to treatment

Other pregnancy outcomes in patients with SSA or SSB antibodies and the absence of NLS.

A small group of studies has looked at pregnancy outcomes (aside from CHB) in SLE patients with SSA antibodies, with conflicting results (Table 2).

Table 2. Summary of studies for SLE patients with SSA antibodies*
Author, year (ref.)ObjectiveStudy designPopulationResults
  • *

    SLE = systemic lupus erythematosus; ACR = American College of Rheumatology; OR = odds ratio; 95% CI = 95% confidence interval.

Ramsey-Goldman et al, 1986 (64)Study of 155 women (331 pregnancies) to look at outcomes for SLE patients with SSA antibodiesRetrospective database studyAll patients with probable or definite SLE according to ACR criteria46% of the pregnancies with SSA and 45% of those without SSA resulted in adverse fetal outcome (prematurity, miscarriage, still birth, and neonatal complication), showing no increased risk associated with the antibody
   47 of 155 women had SSA antibodies and the groups were compared7 of 9 neonatal complications in the SSA group were of cardiac origin and 6 of these 7 were associated with neonatal heart block
    1 pregnancy in the SSA-negative group was associated with a neonatal complication of cardiac origin (endocardial fibroelastosis with heart block)
    A trend was seen associating higher-titer SSA with poor neonatal outcome, although this was not statistically significant
Watson et al, 1986 (65)Study of obstetric histories of 50 SSA-positive women (84 pregnancies) compared with a control group of 47 women (63 pregnancies) with RNP antibodiesRetrospective chart reviewOnly 40% of SSA-positive patients met criteria for SLE, whereas 70% of the patients in the RNP group met criteria for SLEWhen compared with patients with a clear diagnosis of SLE, fetal loss in the SSA-positive group was 45% compared to 21% in the RNP only–positive group. Although this was significant (P = 0.04), the numbers were small.
 A subset analysis looked specifically at SLE outcomes  No difference was seen between the patients without SLE with SSA vs. RNP antibodies
    Black patients with SSA had an increased rate of fetal wastage as compared with those with only RNP (71% vs. 18%; P = 0008)
Petri and Allbritton, 1993 (37)Study of 203 SLE patients and compared outcomes with friends and relatives of the patientsProspectively collected case series481 SLE pregnancies vs. 566 controlsSSA was not associated with pregnancy loss (OR 1.21, 95% CI 0.52–2.84)
   The number of patients with SSA is unclear 
Leu and Lan, 1992 (66)To assess the influence of the effect of SSA antibodies on pregnancy in SLERetrospective chart review with banked sera154 patients with SLE and 123 controlsThose with SSA antibodies and SLE experienced significantly more episodes of intrauterine growth retardation than those without.
   All SLE patients met ACR criteria for SLE3 patients with SSA positivity developed neonatal heart block (0.79%)
   430 SLE gestations with 379 parities 
   67.2% of SLE patients were SSA antibody positive 
Mavragani et al, 1998 (67)Cohort study on patients with autoimmune disease with SSA-positive and SSA-negative statusRetrospectiveOf the patients with autoimmune disease, 78 had positive SSA and 71 did not have SSANo differences in pregnancy outcome or pregnancy loss were detected between SSA-positive and SSA-negative patients (P = 0.49)
   Outcomes were compared between the SLE groups specifically and with 154 healthy controls 
Brucato et al, 2002 (12)Study of 100 patients with autoimmune disease and SSA antibodies and pregnancyProspective and retrospective (looked at prior pregnancies)122 SSA-positive pregnancies and 140 SSA-negative pregnancies were comparedNo differences in pregnancy loss or poor outcomes were detected
   61 SSA-positive SLE pregnancies compared with 86 SSA-negative SLE pregnancies 

Summary.

Available obstetric guidelines (10) identify increased risk in seropositive patients. For example, the European League Against Rheumatism (EULAR) (19) recommendations on SLE identify patients with anti-SSA and anti-SSB antibodies as having a high risk for unfavorable obstetric outcomes, although no clear method of screening is outlined.

Our systematic review did not identify any controlled or noncontrolled studies directly demonstrating that screening for SSA and/or SSB antibodies resulted in improved pregnancy outcomes. However, given the risk of CHB in seropositive mothers, the panel recommended these screening labs be obtained to best determine appropriate monitoring during pregnancy. These laboratory studies should be recorded in the chart in the first trimester and reflect testing completed within the prior 2 years.

QI 1, part 2: aPL screening at pregnancy.

We performed a systematic review examining the predictive value of aPL studies (this would include antibody studies such as anticardiolipin [aCL; IgG and IgM] and β2-glycoprotein [IgG and IgM], as well as functional assays such as Russell's viper venom time or lupus anticoagulant) for pregnant patients with SLE. Our Medline search yielded 878 articles and 583 articles from EMBase. Therefore, we reviewed 1,461 abstracts and chose 52 articles for full-text review. We focused our search on articles that examined SLE patients with aPL without a diagnosis of antiphospholipid antibody syndrome (APS). We excluded studies where patients received anticoagulation. Four articles were deemed appropriate to the research question and used in the final summary below. All were case studies. Only one study controlled for SLE activity when determining the risk of aPL on outcomes. No randomized trials on screening for aPL were found.

Summary of findings.

Tests for aPL, including aCL IgG, aCL IgM, and lupus anticoagulant, are commonly found in the sera of SLE patients (30–40%) (20). Once a test for aPL is repeatedly positive, clinical events determine if the individual meets criteria for APS. Defining clinical events include recurrent first trimester abortion (>3 consecutive losses at less than 10 weeks of gestation), fetal death, severe preeclampsia, or preterm birth due to preeclampsia (21). A positive test for aPL may or may not predict an eventual diagnosis of pregnancy-associated APS, which for patients with SLE correlates with many unfavorable outcomes, including prematurity, intrauterine growth restriction, and fetal demise (14, 22–27). In 1999 and 2006, consensus statements were issued in efforts to define the criteria for APS and pregnancy-associated APS (21, 28, 29).

Some studies indicate an increased risk of poor obstetric outcome for patients with aPL and SLE (Table 3), but this is difficult to separate from pregnancy-associated APS and from SLE itself. Also, the laboratory threshold for the determination of a positive antibody level seems to vary, especially in the older literature. Therapeutic strategies for pregnancy-associated APS are discussed in another QI below.

Table 3. Prognostic studies of pregnant SLE patients with aPL*
Author, year (ref.)ObjectiveStudy designPopulationResults
  • *

    SLE = systemic lupus erythematosus; aPL = antiphospholipid antibodies; ACR = American College of Rheumatology; aCL = anticardiolipin antibodies; APS = antiphospholipid antibody syndrome.

  • All results controlled for SLE activity.

Cortés-Hernández et al, 2002 (23)Observational study of 103 pregnancies in 60 SLE patients with collection of clinical and laboratory dataCase seriesProspective studyIn multiple logistic regression, β2-glycoprotein I antibodies were associated with fetal loss, and those positive for aCL were more likely to experience poor fetal outcomes
   All patients met ACR criteria for SLE 
   40 pregnancies with aCL 
   33 pregnancies with β2-glycoprotein I antibodies 
Johns et al, 1998 (68)Observation study of 54 pregnancies in 28 SLE patientsCase seriesRetrospective studyNo increased risk was noted, but only 26 pregnancies had aPL information and 8 of 26 pregnancies were positive for aPL; of those positive, there were 3 live births and 5 spontaneous abortions (no statistical significance)
   All patients met criteria for SLE 
   26 patients with aPL information 
Lima et al, 1995 (24)Observational study of 108 pregnancies in 90 SLE patients with collection of clinical and laboratory dataCase seriesProspective studyThose who met criteria for APS correlated with fetal loss (P < 0.006)
   All patients met criteria for SLEaCL without a diagnosis of APS correlated with fetal loss (P < 0.1)
    APS did not correlate with preterm birth
Loizou et al, 1988 (69)Retrospective study of 84 women with SLE who had at least 1 pregnancyRetrospective case seriesRetrospective studyFor those with aCL, the percentage associated with fetal loss was 58.7%
   All patients met criteria for SLEIgG aCL presence and level statistically correlated with fetal loss
 143 pregnancies with positive aCL 143 pregnancies with aCLIgM aCL did not show statistical significance with fetal loss
   Results were compared with 26 subjects without SLE with a poor obstetric historyAntibody-negative patients: 27% experienced fetal loss (23 of 93 pregnancies)

Summary.

Guidelines from the Royal College of Obstetrics and Gynecology (RCOG) (30), the American College of Obstetricians and Gynecologists (ACOG) (31), EULAR (19), and the Finnish Medical Society (32) all describe pregnancies with aPL as being linked to poorer outcomes.

The panel believed that, although studies defining the risk of aPL in pregnancy were sparse, the aPL laboratory panel may provide prognostic information that would aid in counseling and determination of appropriate monitoring for pregnant SLE patients. Analogous to SSA and SSB, these antibodies should be noted in the medical record within 4 weeks of documented pregnancy and reflect testing done within the prior 6 months.

QI 2.

IF a patient has had pregnancy complications as a result of APS (including unexplained fetal death after 10 weeks of gestation; birth before 34 weeks as a result of severe preeclampsia, eclampsia, or placental insufficiency; or 3 or more unexplained consecutive spontaneous abortions before the tenth week of gestation in the setting of positive aPL testing, repeatedly positive 12 weeks apart) (29), THEN the patient should be offered aspirin (ASA) and heparin (i.e., heparin or low molecular weight heparin) during subsequent pregnancies, BECAUSE these treatments may improve pregnancy outcomes for patients with APS.

Search strategy/results.

Available trials on anticoagulation for patients with aPL and recurrent pregnancy loss usually excluded SLE patients; therefore, the information summarized in this section stemmed mostly from patients with primary APS. A Cochrane review on this subject was published in 2005, summarizing scientific evidence on this topic published before June 2003. We updated this search to include the period of 2003–2008. We limited our investigation to randomized or quasi-randomized clinical trials for SLE or non-SLE patients during this time period. Two hundred seven articles from Medline and 333 from EMBase were retrieved, totaling 540 citations to be reviewed. Of these references, 25 abstracts were reviewed and 2 full-text articles were included in the summary. The Cochrane review contributed 12 controlled trials (Table 4). The evidence used in this summary includes randomized and quasi-randomized clinical trials as well as a systematic review.

Table 4. Summary of randomized or quasi-randomized trials on treatment for pregnancy-associated APS using heparin, ASA, or combination*
Author, year (ref.)Treatment armsObjectiveStudy designPopulationSLE patients excludedResults
  • *

    APS = antiphospholipid antibody syndrome; ASA = aspirin; SLE = systemic lupus erythematosus; SC = subcutaneously; aPL = antiphospholipid antibodies; OR = odds ratio; 95% CI = 95% confidence interval; IVIG = intravenous immunoglobulin.

Cowchock et al, 1992 (70)Prednisone 20 mg orally twice daily + ASA 80 mg daily vs. heparin 10,000 units SC twice daily + ASA 80 mg dailyTo compare low-dose heparin with prednisone 40 mg daily (both plus low-dose ASA) on live birthsMulticenter randomized trial20 pregnant women with aPL-associated fetal lossNoLive birth rates were the same (75%) with either treatment, but “serious” maternal morbidity and the frequency of preterm delivery were higher among women in the prednisone group (6 of 6 vs. 2 of 8 in the heparin group)
Silver et al, 1993 (71)ASA 81 mg orally daily vs. prednisone orally daily (dose varied) + ASA 81 mg orally dailyTo compare the effect of ASA alone with prednisone plus ASA on morbidity and preterm birthRandomized trial34 pregnant women with aPL-associated fetal lossNoNo perinatal losses were observed; preterm delivery occurred more often in the prednisone + ASA group (8 of 12 vs. 3 of 22 in the ASA-only group)
Kutteh et al, 1996 (72)Heparin 5,000 units SC twice daily + ASA 81 mg orally daily vs. ASA 81 mg orally dailyTo compare the effect of ASA alone with heparin plus ASA on morbidity and preterm birthProspective single-center trial50 women with APS alternately assigned to each treatmentNoLive births occurred in 20 of 25 in the heparin + ASA group and 11 of 25 in the ASA- alone group (P = 0.05); there were no significant differences between the groups
Rai et al, 1997 (44)Heparin 5,000 units SC twice daily + ASA 75 mg orally daily vs. ASA 75 mg orally dailyTo compare the effect of ASA alone with heparin plus ASA on live birthsRandomized trial90 women with recurrent miscarriage and positive aPLYesLive births occurred in 71% (32 of 45 pregnancies) in the heparin + ASA group and 42% (19 of 45 pregnancies) with ASA alone (OR 3.37, 95% CI 1.40–8.10)
Tulppala et al, 1997 (73)ASA 50 mg orally daily vs. placeboTo compare the effect of ASA with placebo on live birthsRandomized trial66 women with recurrent “unexplained miscarriages” (12 with aPL, 54 without aPL)YesOf 58 women with viable pregnancies at 7 weeks, 7 of 33 in the ASA group (4 with aPL and 3 without aPL) and 5 of 33 in the placebo group (2 with aPL and 3 without aPL) experienced miscarriages (not significant); 1 of 33 in the ASA group vs. 3 of 33 in the placebo group experienced preeclampsia (not significant)
Cowchock and Reece, 1997 (74)ASA 81 mg orally daily vs. usual careTo compare the effect of ASA with usual care on live births and low birth weight in low-risk womenRandomized trial19 “low-risk” women positive for aPLYes1 woman in the ASA group had a fetal death and 1 in the usual care group had a low birth weight infant; this difference was not significant
Laskin et al, 1997 (75)Prednisone (0.5 mg/kg) orally daily and ASA 100 mg orally daily vs. placeboTo compare the effect of prednisone and ASA with placebo on live births and prematurityRandomized trial202 women with recurrent fetal loss and at least 1 autoantibody (not necessarily aPL)Yes66 of 101 treatment and 57 of 101 placebo women had live births (P = 0.19). More infants were born prematurely in the treatment group than in the placebo group (62% vs. 12%; P < 0.001). The major side effects of therapy in the mothers were hypertension (treatment group: 13%, placebo group: 5%; P = 0.05) and diabetes mellitus (15% vs. 5%; P = 0.02)
Branch et al, 2000 (76)Heparin 10,000 units SC twice daily, ASA 81 mg orally daily, and 1 gm/kg 10% IVIG × 2 every 4 weeks through 36 weeks of gestation vs. heparin 10,000 units SC twice daily, ASA 81 mg orally daily, and placeboTo compare the effect of heparin, ASA, and IVIG with heparin and ASA only on obstetric and neonatal outcomesRandomized trial16 women with aPL and fetal loss or thromboembolismYesThere were no significant differences between the 2 groups
Pattison et al, 2000 (46)ASA 75 mg orally daily vs. placeboTo compare the effect of ASA vs. placebo on live births in low-risk womenRandomized trial50 women with recurrent miscarriage and aPLYes85% (17 of 20) of the placebo group and 80% (16 of 20) of the ASA-treated group had live infants delivered. This difference was not significant. There were no significant differences in antenatal complications or neonatal morbidity between the groups
Vaquero et al, 2001 (77)IVIG 0.5 gm/kg × 2 monthly through 32 weeks of gestation at center 1 vs. prednisone 10–20 mg orally daily and ASA 100 mg orally dailyTo compare the effect of IVIG with prednisone/ASA on live births and maternal and perinatal morbidity2-center prospective trial82 women with aPL-associated recurrent fetal lossNoLive birth rates were equivalent between groups (78% vs. 76%). Mean birth weight was higher in the IVIG group. In the prednisone/ASA group, gestational hypertension and gestational diabetes mellitus were found significantly more often than in the IVIG-treated group (14% vs. 5% and 14% vs. 5%, respectively)
Farquharson et al, 2002 (41)ASA 75 mg orally daily vs. low molecular weight heparin 5,000 units SC daily + ASA 75 mg orally dailyTo compare the effect of low molecular weight heparin and ASA with ASA alone on live birthsQuasi-randomized trial119 women with recurrent miscarriage and aPL 98 women randomized before 12 weeks of gestationNo34 of 47 women in the ASA group and 40 of 51 in the low molecular weight heparin/ASA group had live births (not significant)
Triolo et al, 2003 (78)IVIG 400 mg/kg/day × 2 monthly through 31 weeks of gestation vs. low molecular weight heparin 5,700 units SC daily and ASA 75 mg orally dailyTo compare the effect of IVIG with heparin and ASA on live birthsRandomized trial40 women with aPL-associated recurrent fetal lossYesThe women treated with heparin plus ASA had a higher rate of live births (84%) than those treated with IVIG (57%; P = 0.06)
Stephenson et al, 2004 (79)Dalteparin + ASA vs. unfractionated heparin + ASATo compare the effect of low molecular weight heparin with unfractionated heparin, started preconception, on live birthsRandomized trial28 women with APSNo1 woman in each group did not conceive; 9 of 13 in the dalteparin group vs. 4 of 13 in the unfractionated heparin group had a successful pregnancy (not significant)
Noble et al, 2005 (42)Enoxaparin 40 mg SC daily + ASA 81 mg orally daily vs. unfractionated heparin 5,000 units SC twice daily + ASA 81 mg orally dailyTo compare the effect of low molecular weight heparin with unfractionated heparin on live births2-center, prospective trial50 women with ≥3 pregnancy losses and positive aPLNo21 of 25 women in the low molecular weight heparin group and 20 of 25 in the unfractionated heparin group delivered a viable infant (these differences were not statistically significant)

Summary of evidence.

Pregnancy-associated APS is defined as aPL abnormalities accompanied by poor obstetric outcomes (21). Approximately 30–40% of patients with SLE express these antibodies in the serum. The risk of future miscarriage in a seropositive woman, not specific to SLE, with a history of 3 recurrent spontaneous abortions is estimated at 85% (33, 34), indicating a need for therapeutic intervention in these patients.

Obstetric outcomes of pregnant patients with aPL antibodies.

Multiple studies document the association between APS and poor obstetric outcomes for patients with SLE (24, 35, 36), and both conditions (APS and SLE) confer independent risk for unfavorable outcomes in pregnancy (37, 38). Although aPL are common in women with SLE, to our knowledge, no randomized controlled trials exist to establish the benefit of treating pregnancy-associated APS in this specific population. For many of the clinical trials on pregnancy-associated APS therapy, SLE patients have been excluded, perhaps to separate the independent risk of SLE alone on outcomes (Table 4).

Two systematic reviews, one in 2005 and one in 2006, looked at treatment of pregnancy-associated APS, and both noted the superiority of heparin plus low-dose ASA for prophylaxis (39, 40). A recent Cochrane review found that unfractionated heparin plus ASA was superior to ASA alone (relative risk [RR] 0.46, 95% confidence interval [95% CI] 0.29–0.71) (40).

Low molecular weight heparin, commonly used in clinical practice for its convenience, has been less well studied. Farquharson et al found no difference between low molecular weight heparin and ASA alone in one prospective trial of 98 patients (RR 0.78, 95% CI 0.39–1.57) (41); however, another study found similar live birth rates when comparing low molecular weight heparin plus ASA versus unfractionated heparin plus ASA (42).

In perhaps the 2 most notable trials showing the benefit of heparin plus ASA, most patients met pregnancy-associated APS criteria because of first trimester pregnancy losses and with low-titer antibody studies, perhaps indicating a milder spectrum of disease (43, 44). It should also be noted that the effect of ASA alone varied greatly between studies, perhaps indicating some of the weaknesses in generalizability and in study design. In some trials, patients receiving ASA alone reported an ∼80% live birth rate (45, 46), whereas others reported an ∼40% live birth rate (43, 44).

Summary.

Many current guidelines, such as those from the ACOG, support the use of heparin and ASA for pregnancy-associated APS (31). In 2003, the RCOG stated that for “women with a history of recurrent miscarriage and aPL, future live birth rate is significantly improved when a combination therapy of ASA plus heparin is prescribed” (30).

For patients with a history of documented pregnancy-associated APS, the risk of poor obstetric outcomes is significant, and evidence suggests that prophylaxis may improve outcomes in subsequent pregnancies. Although there may be some benefit for the use of ASA alone in these patients, the panel believed that the combination of heparin (either low molecular weight heparin or unfractionated heparin) plus low-dose ASA provides superior results in increasing the live birth rate.

QI 3.

IF a woman between 18 and 45 years of age is started on a medication for SLE (e.g., chloroquine, quinacrine, methotrexate, azathioprine, leflunomide, mycophenolate mofetil, cyclosporine, cyclophosphamide, or thalidomide), THEN a discussion with the patient about the potential teratogenic risks of therapy and about contraception should be documented prior to drug initiation, unless the patient is unable to conceive (e.g., the patient had a hysterectomy, oophorectomy, or tubal ligation, or is postmenopausal), BECAUSE these drugs either have teratogenic potential or pose an unknown risk to the developing fetus.

Search results.

The search yielded 171 articles from Medline and 414 articles in EMBase; therefore, 585 abstracts were reviewed. Of these, 10 articles were screened in full-text format and no relevant articles were used in the final summary. We did find 3 applicable articles from hand searching and discussion with experts in the field. These references are cited in the summary below. All of the evidence was derived from observational data.

Of note, the panel chose not to include nonsteroidal medications or hydroxychloroquine in this QI so as not to indicate that these medications should be avoided in pregnancy (47–53). Other drugs were carefully considered. For example, multiple studies, especially from renal transplant and inflammatory bowel literature (54, 55), support the relative safety of azathioprine and prednisone during pregnancy, if needed. By including these medicines in the QI, the panel believed that counseling about potential risks and benefits was still warranted so that the patient could make an informed decision regarding future pregnancies.

Summary.

SLE often affects women during the childbearing years; therefore, the desire for pregnancy may coincide with the use of immunosuppressive medications. The decision to discontinue medications prior to conception requires consideration of the woman's clinical state and the drug's potential fetal toxicity. Unfortunately, fewer than 20% of women receiving teratogenic drugs receive appropriate counseling (56–58), and approximately 6% of pregnancies in the US are exposed to potentially teratogenic class D or X medications (59), suggesting deficits in the quality of care in this area. No trial data exist to identify if preconception counseling for SLE patients results in a lower incidence of unplanned pregnancies or improves pregnancy outcomes.

In the general population, only a few observational studies have attempted to study the effects of contraceptive counseling for patients receiving teratogenic drugs. In one managed care cohort followed by Schwarz et al, 1 of 6 women had filled a prescription for a Food and Drug Administration (FDA) class D or X drug within the previous year. The incidence of filled prescriptions followed by a documented positive pregnancy test was similar among women taking FDA class A and B medications as compared with women taking class D or X medications. This perhaps indirectly indicates a failure or lack of counseling for those patients receiving drugs with high teratogenic potential (58).

In another large observational study, Weisman et al interviewed 898 women of childbearing age without lupus (SLE) and found that a history of contraceptive counseling correlated with a decrease in unintended pregnancy and a higher likelihood of contraception use (60).

Patients with SLE receive many medications with either teratogenic potential (Table 5) or unknown effects on the developing fetus. It is not clear if preconception counseling decreases the rate of unintended pregnancy in SLE, but patients in the childbearing years should be made aware of the potential teratogenic risks of therapy in order to best promote informed decisions. In a current QI set for rheumatoid arthritis, documentation of counseling is recommended when starting a drug of teratogenic potential (1). The panel recommended documentation of this counseling in the medical chart.

Table 5. US FDA pregnancy risk categories for common medications used for systemic lupus erythematosus*
DrugFDA ratingPossible effects on fetus
  • *

    Adapted, with permission, from ref.1 (Seminars in Arthritis and Rheumatism, Vol. 35, Khanna D, Arnold EL, Pencharz JN, Grossman JM, Traina SB, Lal A, et al, Measuring process of arthritis care: the Arthritis Foundation's quality indicator set for rheumatoid arthritis, p. 211–37, 2006, with permission from Elsevier). FDA = Food and Drug Administration; category C = risk cannot be ruled out, human studies are lacking; category X = contraindicated in pregnancy; category D = positive evidence for risk to human fetus; category B = no risk to human fetus despite possible animal risk or no risk in animal studies, human studies have not been done (not shown); category A = controlled studies show no fetal risk (not shown).

Chloroquine/quinacrineCPossible ear and eye toxicity
MethotrexateXCraniofacial and extremity defects
LeflunomideXEmbryotoxic
AzathioprineDFetal growth retardation, decreased thymic shadow, adrenal hypoplasia, depressed immunoglobulin levels, chromosomal abnormalities
CyclophosphamideDFetal growth retardation, bone marrow suppression, infection, hemorrhage, chromosomal abnormalities, developmental delay
CyclosporineCPremature birth, low birth weight
Mycophenolate mofetilDCraniofacial abnormalities, extremity defects, other fetal malformations
ThalidomideXExtremity defects

DISCUSSION

QIs have emerged as helpful tools in the assessment of the delivery of care. Using a validated approach, we have developed a QI set for SLE. The reproductive health QIs in our set encompass 3 important topics, including screening for autoantibodies, treatment of pregnancy-associated APS, and counseling regarding risk and contraception in women taking potentially teratogenic medications. In this study, we have outlined the scientific evidence supporting these reproductive health indicators, detailing both our systemic literature reviews and the results of our expert panel process. Establishment of QIs provides a framework to examine current processes of care and may aid in the improvement of health care delivery.

Although we used a validated approach to develop the reproductive indicators, several caveats should be taken into consideration. As noted in our summary, studies specific to SLE are limited, necessitating that in some cases, we extrapolated from studies in non-SLE subjects. Also, the current consensus statement for pregnancy-associated APS was not applied to all cited studies in the systematic review, perhaps affecting generalizability. Finally, we acknowledge that QIs are limited in scope and do not cover many important aspects of care pertaining to reproductive health in SLE. Unlike guidelines, which attempt to provide prospective guidance on optimal patient care, QIs instead allow the retrospective assessment of certain limited measurable aspects of care.

In summary, we present 3 reproductive health QIs for the care of patients with SLE. These indicators were derived from a validated process using a systematic literature search combined with the clinical opinion of experts in the field. In the future, we hope to use these indicators to better assess the quality of care provided in SLE and to guide improvements in health care delivery to this population.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Drs. Gillis and Yazdany had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Gillis, Panopalis, Yazdany.

Acquisition of data. Gillis, Schmajuk, Yazdany.

Analysis and interpretation of data. Gillis, Ramsey-Goldman, Yazdany.

Acknowledgements

The authors wish to thank the participating librarians, including Gloria Won at the University of California, Kimberly Schwartz at Stanford University, and Rosalind Dudden at National Jewish Health. In addition, the authors thank Caroline Gordon, MD, and Virginia Winn, MD, PhD, for their review of draft indicators.

Ancillary