Association between pregnancy outcomes and cardiovascular deaths in women with systemic lupus erythematosus utilising Swedish population registries
Background and aim
Survival of patients with systemic lupus erythematosus (SLE) has improved with the advent of more effective treatment, and cardiovascular mortality is now the leading cause of death in this group of patients accounting for approximately 30 – 50% of all deaths.(18), (122) A meta-analysis on cardiovascular death in patients with SLE shows that the risk of cardiovascular death was increased three-fold compared to the general population (meta-standardized mortality ratio 2.72, 95% CI 1.83-4.04).(19) Young women with SLE (age 35-44 years) are particularly vulnerable.
Women with SLE are at increased risk of adverse pregnancy outcomes related to placental disease. They have a much higher risk of developing pre-eclampsia (14-23%), and fetal growth restriction (20-30%). These complications are also associated with the higher rate of pre-term (< 37 weeks gestation) delivery (20-30%) observed in this group of women. (162), (161) These complications are related to placental disease and are collectively known as maternal-placental syndrome (MPS).(142)
In unselected populations, the risk of cardiovascular death is increased by two-fold in women whose pregnancies have been complicated by pre-eclampsia a feature of MPS.(140), (203) What remains unclear is whether the higher prevalence of cardiovascular death observed in women with SLE is associated with history of MPS, or whether these cardiovascular deaths occur independent of an adverse obstetric history.
The aim of this study was to explore the association between pregnancy complications related to MPS and the future risk of cardiovascular death in parous women with SLE.
This is a population-based retrospective study of women with SLE in Sweden between 1973 and 2011. Women with SLE were identified from the National Patient Register (NPR) and the Hospital-based Outpatient Care Register. These databases were then linked using the personal identity number (PIN) to the Swedish Medical Birth Registry (MBR) and Cause of Death Registry to obtain information on pregnancies and cause of death during the period of study.
Compulsory reporting of all inpatient admissions was established from 1987, with more than 99% of all hospital discharges now included in The National Patient Register (NPR). (176)Diagnoses made during each admission were coded using the Swedish international classification of disease (ICD) system.(176) The ICD codes used during the study period ranged between ICD 8 to ICD 10.The recording of diagnostic codes is well validated, specific examples include a more than >90% positive predictive value (PPV) for coronary artery disease and > 85% PPV for rheumatoid arthritis. (176) In 2001, the Hospital-based Outpatient Care Register was established, but its coverage was only 80% of all outpatient attendees. (176) The Medical Birth Register (MBR) founded in 1973, includes data on 97.0-99.5% all pregnancies and deliveries beyond 22 weeks in Sweden. The Cause of Death Register contains information on all deaths of Swedish nationals irrespective of whether the death occurred in Sweden or elsewhere. (http://www.socialstyrelsen.se/ statistics).
We used median and interquartile ranges to summarise continuous variables and the Mann-Whitney U test for comparisons between groups. We used 2 test for univariate comparisons of dichotomous data. The risk of an event was modelled using logistic regression. Univariate analysis was performed on each of the outcomes of interest. Multivariate analysis was performed adjusting for the year of first delivery to account for changes in diagnostic thresholds, coding of disease and management of co-morbidities leading to cardiovascular death (the cohort effect). We also adjusted for number of inpatient admissions, and cardiovascular risk factors (renal disease, diabetes and hypertension). Interaction tests were performed using the Wald test to explore the effect of SLE at the time of pregnancy on the association between MPS and future cardiovascular death. All p-values were two sided and the significance was set at p<0.05. Analyses were performed with Stata IC version 13.1. (StataCorp LP, College Station, TX).
Between 1973 and 2011 there were 12,702 women in Sweden with SLE identified from the NPR. In this cohort, there were 7,410 pregnancies in 3,977 women. All descriptive and analytical data relate to women with history of pregnancy (n=3,977). In this group 325 (8.2%) women had died. The commonest primary cause of death was malignancy (32.3%, n= 105), followed by SLE or connective tissue-related deaths 26.8% (n=87). Primary cardiovascular deaths from coronary artery disease (n=31), stroke (n=11) and peripheral vascular disease (n=2) was the third commonest primary cause of death and accounted for 13.5% (n=44) of all deaths. In addition, cardiovascular disease was recorded as a contributory cause in 32.3% (n=105) of all deaths in parous women with SLE.
The median age of cardiovascular death was 54 years (IQR 48-58) and not significantly different from the age of the women who had died from other causes [53 years (IQR 45-60)] (p=0.52). The median age of women alive in study cohort is 50 years (IQR 41-59). Compared to women who were alive, there was no difference in the duration of SLE or their age of diagnosis of SLE amongst those who had died from cardiovascular causes. However, women who died from cardiovascular causes had been admitted more frequently to hospital. The cardiovascular risk factors such as renal disease, hypertension and infection were higher in those who died from cardiovascular death. (Table 3.1)
When pregnant, maternal age, maternal SLE in pregnancy, parity and BMI were similar between those who were alive and those who had died from cardiovascular disease. However, compared to women who were alive, women who died from cardiovascular diseases were more likely to smoke in pregnancy, had lower livebirth rates, infants with lower birthweights and were also more likely to have had a preterm delivery before 34 weeks.
The prevalence of MPS in the study cohort was 22.1% (n=811). There were overlaps between the individual features of MPS i.e. there were 401 women with recorded history of hypertensive disorder of pregnancy that was the sole feature of MPS in 65%. Of these women 29.4% had concurrent SGA, and 4.5% had also had stillbirth and 1.0% placental abruption. (Figure 3.1) The overall prevalence of MPS was higher in women who died from CV death (34.1% vs. 21.9%, p=0.05). (Table 3.1)
A history of MPS in any pregnancy was associated with a two-fold increased risk of cardiovascular death (Adjusted (adj) OR 2.19: 95% CI 1.14 – 4.22) (Table 3.2). Placental abruption was the only feature of MPS specifically associated with risk of primary cardiovascular death (Adj OR 5.78; 95% CI 1.61 – 20.72). Preterm delivery < 34 weeks independently (Adj OR 2.22; 95% CI 1.03 – 4.79) or in association with MPS (Adj OR 2.49; 95% CI 1.06 – 5.85) were also associated with an increase in risk of cardiovascular death. These associations described were independent of whether maternal SLE was present at the time of pregnancy or not (p>0.05 for all interaction tests performed).
Discussion and conclusion
In our cohort of parous women with SLE, the risk of subsequent cardiovascular death was increased in those who had had a pregnancy complicated by MPS. This association was independent of any cohort effect, medical comorbidities, and cardiovascular risk factors. The risk of cardiovascular death was highest in those who had a history of placental abruption complicating their pregnancy.
The underlying pathophysiology for accelerated cardiovascular disease in young women with SLE is unclear. (19) While most patients with SLE do have traditional risk factors for vascular disease, this still fails to account for the atherosclerotic disease burden seen in this group. There are conflicting reports of the contribution of lupus activity, anti-phospholipid syndrome, disease duration and corticosteroid as non-traditional risk factors for cardiovascular disease in patients with SLE.(94), (204), (48) Multiple large international studies have been unsuccessful in ascertaining the cause for excess cardiovascular disease even when taking into account SLE-specific disease damage and drugs. (185)
In unselected populations, the relationship of adverse pregnancy outcomes – specifically, hypertensive disorders of pregnancy – and cardiovascular death causes is well established. More recent meta-analyses have confirmed this, though the nature of this association, and its exact pathophysiology remains obscure. (140), (203) Hypertensive disorders of pregnancy are a result of abnormal placental vasculature. This abnormal placentation manifests itself clinically by pre-eclampsia, fetal growth restriction or the delivery of a SGA infant, placental abruption or stillbirth. (142) These conditions are collectively known as MPS. It is postulated that there could be shared pathophysiological processes between MPS and future cardiovascular disease that are unmasked by the stress of pregnancy, or pre-existing cardiovascular risk factors (i.e. hypertriglyceridaemia, hypertension, etc.) that are present in young women who then develop MPS during pregnancy.
Whilst our study did not demonstrate an effect of presence of maternal SLE on the association between MPS and future cardiovascular death, it is well established there is often a latency between clinical features of SLE and need for hospital admission. A similar diagnostic-lag has also been demonstrated in the parous SLE cohort. (205) Pre-eclampsia and placental abruption is associated with elevation of soluble months post-partum and, it has been hypothesized, may contribute to persistent endothelial dysfunction and thereby progressive atherosclerosis. The placental synthesis of these anti-angiogenic factors may also be higher in women with SLE who already have a predisposition to vascular disease. As in the general population, sFLT-1 may be a useful marker for pre-eclampsia in women with SLE and does persist in the circulation after delivery. (172)
In women with early onset pre-eclampsia there is an excess shedding of placental microparticles (that can cause apoptosis, angiogenesis, inflammation and vascular dysfunction), resulting in a proinflammatory response by activating mononuclear cells and neutrophils. (150) In our study, placental abruption has been identified as the strongest risk factor for future cardiovascular disease. In the setting of SLE (where immune dysregulation including B-cell activation and defective T-cell responses drive the inflammatory response and perpetuate the continuous cycle of immune complex deposition), the added challenge of placental microparticles and other inflammatory molecules could stimulate the accelerated atherosclerosis in young women. In other words, inflammation as a result of placental dysfunction could work synergistically with the inflammatory milieu of SLE to accelerate the development of future cardiovascular disease.
The incidence of preterm deliveries has increased worldwide. This is in part because older women with multiple medical comorbidities are now attempting pregnancy, an increasing incidence of obesity and also increased iatrogenic preterm births secondary to changes in obstetric practices. In our study, length of gestation was very well captured in the MBR. As a whole, preterm delivery 34 weeks gestation was associated with an increased risk of subsequent cardiovascular death. In this study, we chose 34 weeks as a cut-off to define preterm delivery as the most related severe placental-related iatrogenic deliveries are performed before 34 weeks. We do not have information on the indication of iatrogenic pre-term delivery, however as length of gestation (and consequently preterm deliveries) is well captured in the MBR, it serves as a useful marker for identifying women at greater risk of future cardiovascular disease.
The strengths of our study were a large population-based cohort of 12,702 Swedish women with SLE over a 38-year period. However, as MBR data were only available from 1973 onwards, this was a relatively young parous group (median age of 50 years), and there were only 44 women who had a history of pregnancy and had died from cardiovascular causes. Despite the women being relatively young, there were still 325 (8.17%) deaths. In this analysis, we deliberately excluded non-parous women with SLE as they are a heterogeneous group. The non-parous women may have made a personal choice not to become pregnant, or were unable to do so because of more severe disease, or been rendered infertile as a result of cytotoxic drug use (especially cyclophosphamide), or suffered recurrent early miscarriages due to severe or active disease. All these factors could act as potential confounders.
Other limitations included the lack of information pharmacologic data about steroid and hydroxychloroquine use, disease activity and damage, socioeconomic status and concurrent antiphospholipid syndrome. Information on disease activity and damage is particularly difficult to collect from retrospective population databases. We could have used presence of renal disease and multiple admissions to hospital as surrogate markers for more severe SLE, however, further risk stratification of the group would make data interpretation difficult. Data on the NPR are limited for smoking, cardiovascular risk factors such as dyslipidaemia and family history of premature cardiovascular disease (<60 years). (206) Information on maternal smoking and obesity were available from 1982; given the significant amounts of missing data, we did not adjust for it in the multivariate analysis.
Our study demonstrates that in this high-risk population of young women with SLE, adverse obstetric outcomes related to MPS remains an independent predictor of cardiovascular mortality. In an unselected population, a history of pre-eclampsia, a feature of MPS is now included in the risk assessment of cardiovascular disease in women. (207) In women with SLE who have a higher background risk of cardiovascular disease, clinicians may wish to consider targeted screening for cardiovascular disease in women with previous pregnancies complicated by MPS, particularly if they had placental abruption complicating their pregnancies. This should be considered regardless of whether there was a history of maternal SLE at the time of the pregnancy.
Table of contents
Table of contents
List of Tables
List of Figures
1. Co-Authorship Chapter 1
2. Co-Authorship Chapter 3
3. Co-Authorship Chapter 4
4. Co-Authorship Chapter 5
1.1 Introduction to Systemic Lupus Erythematosus (SLE)
1.2 SLE in New Zealand
1.3 Morbidity and mortality in SLE
1.4 Cardiovascular events in SLE
1.5 Traditional cardiovascular risk factors
1.6 SLE-specific risk factors
1.7 The role of gender
1.8 The association between maternal-placental syndrome and cardiovascular events in the
1.9 Possible hypothesis and future directions in research
2.1 Study design
2.2 Ethical approval
2.3 Study population
2.4 Swedish population registries
2.5 Plan of analysis
2.6 Statistical methods
2.7 Brief overview of the linked population database
2.8 In conclusion
3. Association between pregnancy outcomes and cardiovascular deaths in women with systemic lupus erythematosus utilising Swedish population registries
3.1 Background and aim
3.4 Discussion and conclusion
4. Maternal-Placental Syndrome and Future Risk of Accelerated Cardiovascular Events in Parous Swedish Women with Systemic Lupus Erythematosus
4.5 Key Messages
5. Can differences in gestational age at preterm delivery assist in maternal cardiovascular risk stratification in parous women with SLE?
5.4 Conclusions and discussion
6. Conclusions, discussion and future directions
6.1 Summary of the main findings
6.2 How these study findings may influence clinical practice
6.3 What advice about risks for future cardiovascular health should be shared with a woman with SLE planning pregnancy?
6.4 Utilising a population registry for this analysis
6.5 The strengths of population registries
6.6 Limitations of using population registries in this analysis
6.7 Future directions
GET THE COMPLETE PROJECT
Are adverse pregnancy outcomes associated with the accelerated development of cardiovascular events in parous women with systemic lupus erythematosus?