Healthcare costs of pregnancy in systemic lupus erythematosus: retrospective observational analysis from a US health claims database

Abstract

Objective:

Systemic lupus erythematosus is a complex autoimmune disease, most frequently affecting women of childbearing age. Women with lupus are at increased risk of pregnancy complications that are exacerbated by active disease. Despite this, their use of medications and hospital resources has not been extensively studied.

Methods:

Retrospective analyses of the Truven Health MarketScan database (2006−2012) aimed to quantify drug and resource utilization in pregnant women with lupus, as well as the incidence of pregnancy complications in these patients. Records of women aged 12−54 were reviewed and both lupus patients and pregnancies identified. Pregnant women with lupus were matched 1:5 with either pregnant women without lupus, or non-pregnant women with lupus.

Results:

Pregnancies with lupus were associated with increased complications when compared to pregnancies without lupus. During pregnancy, the use of immunosuppressants decreased in pregnant women with lupus, as did rheumatologist visits, while the number of women not treated with any immunosuppressant increased. Pregnant women with lupus showed higher overall treatment costs than controls. However, compared to non-pregnant women with lupus, medication costs actually dropped, possibly due to the withdrawal of medications from these patients or women becoming pregnant while disease activity was low.

Conclusions:

The large database analyses reported here revealed that pregnancies in women with lupus were associated with a higher risk of complications, higher healthcare costs, and fewer prescribed medications, including immunosuppressants, than the control groups. The increased risk of complications and decreased immunosuppressant use suggest that patients require additional guidance from physicians to give them the best chance of experiencing a safe pregnancy. Indeed, despite the recognized role active lupus plays in increasing pregnancy complications, women with lupus had fewer rheumatology visits during pregnancy, although their visits to their general practitioner/primary healthcare provider increased, highlighting the need for team-based co-ordination care between OBGYN physicians and rheumatologists.

Keywords::

• Systemic lupus erythematosus
• Pregnancy
• Healthcare costs

Introduction

Systemic lupus erythematosus most frequently affects women of childbearing age, with ∼90% of lupus patients being women, most commonly between 15–45 years of age1^,2.

The effect of pregnancy on lupus disease activity is still controversial, with most reports describing increased periods of high disease activity, although some studies report no such change3–6. Lupus nephritis, which may occur for the first time or worsen during pregnancy, can be associated with maternal hypertension and pre-term birth7. Pregnant women with lupus have an increased risk of complications, including spontaneous abortion, fetal growth restriction, pre-eclampsia, eclampsia, and premature rupture of membranes, which may lead to pre-term delivery and/or maternal mortality6^,8–10.

The occurrence of complications correlates strongly with active disease during pregnancy, indicating a particular need to control disease activity during pregnancy4^,11. Although the manifestations of lupus can be controlled by medication, there are concerns about the use of currently available drugs during pregnancy, as some drugs used to treat lupus and its associated co-morbidities can be teratogenic7^,11^,12.

Historically, healthcare professionals advised women with lupus against pregnancy, due to their increased risk of complications. However, with appropriate care and planning it is entirely possible for women with lupus to experience safe pregnancies and have healthy babies4. In particular, the identification of risk factors, such as active disease, and the taking of steps to combat these prior to conception and during pregnancy, are helpful to ensure a successful pregnancy.

Lupus is associated with high healthcare costs and has a considerable impact on healthcare systems13^,14. Additionally, maternal health services represent a substantial portion of hospital expenditures15. Despite this, the use of medications and the utilization of hospital resources in pregnant women with lupus has not been extensively studied.

Furthermore, recent research has shown that women with chronic inflammatory diseases, such as lupus, do not feel that the current information available to them regarding pregnancy is adequate16. For example, one study showed that one third of women with Inflammatory Bowel Disease believed that all medications should be stopped once they become pregnant17. It is particularly important that information surrounding appropriate medication use is made available to women during early discussions regarding their disease, given the large proportion of pregnancies that are unplanned18.

This study used the Truven Health MarketScan database, a large US health claims database, to: (i) consider resource utilization in pregnant women with lupus compared to pregnant women without lupus and non-pregnant women with lupus; (ii) compare the drug utilization characteristics between pregnant women with lupus and non-pregnant women with lupus; and (iii) quantify the reported incidence of maternal and fetal complications in pregnant women with lupus, compared to pregnant women without lupus.

Methods

Data

Data were derived from two retrospective observational analyses of the Truven Health MarketScan US administrative healthcare claims database from 2006–2012. This database is Health Insurance Portability and Accountability Act (HIPAA) compliant and all patient data were de-identified prior to analysis. This study was designed and monitored in accordance with the ethical principles of Good Pharmacoepidemiology Practice (GPP).

Patients

These database analyses reviewed the records of women aged 12–54 years, comparing pregnant women with lupus with either pregnant women without lupus or non-pregnant women with lupus.

Pregnancies were identified by including codes for live birth, miscarriage, and first or second trimester ultrasound. Patients lacking these codes throughout the study period were classified as not pregnant. Lupus patients were identified by ICD-9 codes for 710.0 on at least 2 different days, one of which must have occurred before the beginning of the study period.

For inclusion in these studies, individuals must have been continuously enrolled during the 180-day pre-index period and the 299-day study period. The index date was the date of conception, calculated as a number of days prior to the code that identified them as pregnant. This was 67 days for a miscarriage code, 79 days for a first-trimester ultrasound, 158 days for a second-trimester ultrasound, and 271 days for a birth. Where patients had more than one of these codes available, the earliest code was used to calculate the index date. For non-pregnant women with lupus, an equivalent index date was used: the most recent 710.0 ICD-9 code that permitted a post-index study period of 299 days.

For these analyses, pregnant women with lupus were matched 1:5 with an appropriate control group: either pregnant women without lupus, or non-pregnant women with lupus. Women were matched by index year and quarter, age, geography, data type (fee for service/encounter/medicare/medicare encounter), insurance plan type, and employment status (employed/unemployed).

Variables analyzed

In the first database analysis, the exploratory variables compared between pregnant women with lupus and pregnant women without lupus included resource use during pregnancy and the post-natal period, prior codes for birth, contraception, miscarriage, complications or hospitalizations, fetal complications (including disproportion, abnormality of organs and soft tissues of pelvis, known or suspected fetal abnormality affecting management of the mother, other known or suspected fetal and placental problems affecting management of the mother, polyhydramnios, other problems associated with amniotic cavity and membranes, abnormality in fetal heart rate or rhythm, absence or hypoplasia of umbilical artery), and the total all-cause healthcare costs during pregnancy and the post-natal period, adjusted to the 2012 Consumer Price Index. All-cause healthcare costs included inpatient costs (surgical, medical, obstetric, and other inpatient services costs), outpatient costs (emergency department, office, obstetric, laboratory, pathology, radiology, and other outpatient services costs), and medication costs (for immunosuppressive and non-immunosuppressive medications).

For the second database analysis, the exploratory variables compared between pregnant women with lupus and non-pregnant women with lupus included general practitioner (GP)/primary healthcare provider (pHCP) visits, specialist and hospitalization visits, therapeutic treatments, and the total all-cause healthcare costs.

Statistical analyses

All data were analyzed using SAS statistical software (Version 9.3). Descriptive statistics (mean, standard deviation, median) were employed for continuous variables, while categorical variables were described as the percentage of patients in each group. Conditional logistic regression was used to calculate p-values. These statistical tests were not corrected for multiple comparisons.

Results

Patient populations

In the first database analysis, 1721 pregnant women with lupus were matched with 8605 pregnant women without lupus. In the second analysis it was only possible to match 1634 of the 1721 pregnant women with lupus in the database 1:5 with non-pregnant women with lupus. Therefore, this analysis consisted of 1634 pregnant women with lupus, matched with 8170 non-pregnant women with lupus. Baseline characteristics for women included in these analyses are shown in Supplemental Table 1.

Table 1. Patients’ prior medical history.

	Pregnant women with lupus (n = 1721)	Pregnant women without lupus (n = 8605)	p-value*
Prior birth (%)	22.6	21.6	0.3271
Prior miscarriage (%)	13.7	11.5	0.0098
History of pre-term labor (%)	4.2	2.1	<0.0001
Number of days since last contraception code, mean (SD)	367.7 (370.4)	302.6 (327.4)	0.0003

*p-value were calculated using conditional logistic regression.

SD, standard deviation.

Pregnant women with lupus compared to pregnant women without lupus

Costs of healthcare during pregnancy

For pregnant women with lupus compared to pregnant women without lupus, the mean (SD) total all-cause direct healthcare costs during the 44-week pregnancy and post-partum period were $21,509 ($24,438) and $11,481 ($10,619; p < 0.0001), respectively (Figure 1a; Supplemental Table 2A). Costs were higher in all sub-categories, including the cost of medication and both outpatient and inpatient costs.

Figure 1. Healthcare costs associated with lupus treatment and pregnancy for: (a) Pregnant women with lupus compared to pregnant women without lupus; (b) Pregnant women with lupus compared to non-pregnant women with lupus. **p < 0.0001. Data shown are mean values.

Resource utilization during pregnancy

Pregnant women with lupus had a significantly longer length of stay in the maternity ward compared to pregnant women without lupus (mean [SD] = 2.8 [5.3] days and 1.8 [3.2] days, respectively; p < 0.0001) (Figure 2a, Supplemental Table 3A). They also required a higher number of obstetrics and gynecology (OBGYN) visits, fetal ultrasounds, fetal echocardiographies, and fetal stress and non-stress tests, compared to pregnant women without lupus (Figure 2a).

Figure 2. Resource utilization in: (a) Pregnant women with lupus compared with pregnant women without lupus; (b) Pregnant women with lupus compared with non-pregnant women with lupus. **p < 0.0001. Data shown are mean (SD). CVS, Chorionic villus sampling; GP, General practitioner; OBGYN, Obstetrics and gynecology; pHCP, Primary healthcare provider.

Medical history prior to current pregnancy

There was no significant difference between the proportions of pregnant patients with or without lupus who had had a prior birth (Table 1). However, pregnant women with lupus had a significantly higher number of mean days since their last contraception code, and a significantly higher proportion were associated with a prior miscarriage code and a history of pre-term labor (Table 1).

Pregnancy complications in pregnant women with lupus and pregnant women without lupus

Compared to pregnant women without lupus, pregnant women with lupus had an increased risk of threatened miscarriage (i.e., vaginal bleeding occurring within the first 20 weeks of pregnancy) or premature labor without delivery (i.e., initiation of labor [contractions, cervical dilation] without subsequent delivery because labor symptoms either stopped spontaneously or were halted medically), hypertension, pre-eclampsia, pre-term delivery, and stillbirths (Figure 3).

Figure 3. Complications in pregnancy over the study period. **p < 0.0001, *p < 0.05. Data shown are the percentage of patients experiencing specified complication.

Overall, there was a 7.5% increase in the risk of fetal complications in pregnant women with lupus compared to those without. A known or suspected fetal abnormality that affected the management of the mother occurred in 39.7% of pregnant women with lupus compared to 28.0% of pregnant women without lupus (p < 0.0001).

Pregnant women with lupus compared to non-pregnant women with lupus

Costs of healthcare during pregnancy

When the mean (SD) total all-cause direct healthcare costs were compared between pregnant women with lupus and non-pregnant women with lupus, the costs were significantly (p < 0.0001) higher in pregnant women with lupus ($20,665 [$20,709]) compared to non-pregnant women with lupus ($12,591 [$27,896]) (Figure 1b, Supplemental Table 2B). This was due to differences in the amount of care received, rather than medication costs, as the cost of medication was actually less in pregnant women with lupus ($1915 [$5365]) than non-pregnant women with lupus ($2501 [$5870]).

Resource utilization during pregnancy

In comparison to non-pregnant women with lupus, pregnant women with lupus had a significantly higher number of GP/pHCP and outpatient hospitalization visits (Figure 2b, Supplemental Table 3B).Importantly, pregnant women with lupus had significantly fewer visits to their rheumatologists or internal medicine care providers (Figure 2b, Supplemental Table 3B).

Prescribed drugs used during pregnancy

Compared to non-pregnant women with lupus, significantly more pregnant women with lupus were treated with non-steroidal anti-inflammatory drugs (NSAIDs) (24.1% and 28.9%, respectively; p < 0.0001) (Figure 4). For those patients prescribed corticosteroids, the daily corticosteroid dose was not significantly different between pregnant and non-pregnant women with lupus (respective mean [SD] daily doses [mg/day] were: Hydrocortisone: 28.1 [28.7] vs 32.5 [17.9]; Methylprednisolone: 12.2 [6.2] vs 13.0 [5.9]; Prednisone: 14.6 [14.8] vs 14.7 [15.5]).

Figure 4. Medications prescribed during the study period. **p < 0.0001. Data shown are the percentage of patients receiving specified medication. NSAIDs, Non-steroidal anti-inflammatory drugs.

All immunosuppressant use was significantly higher in non-pregnant women with lupus compared to pregnant women with lupus (Figure 4). Pregnant women with lupus were given fewer immunosuppressants with an FDA pregnancy rating of C, D or X, whilst there was no change in the percentage of women receiving drugs with a B rating (for definitions see Supplemental Table 4). Overall, there was a highly significant increase in the number of pregnant women with lupus receiving none of these treatments compared to non-pregnant women with lupus (52.0% and 33.0%, respectively; p < 0.0001) (Table 2).

Table 2. Immunomodulators and immunosuppressants, by FDA category, prescribed to pregnant women with lupus and non-pregnant women with lupus.

FDA pregnancy category*	Pregnant women with lupus (n = 1634)† n (%)	Non-pregnant women with lupus (n = 8170)† n (%)	p-value‡
B	16 (1.0)	104 (1.3)	0.31
C	719 (44.0)	4684 (57.3)	<0.0001
D	169 (10.3)	2458 (30.1)	<0.0001
X	30 (1.8)	883 (10.8)	<0.0001
None of the above	850 (52.0)	2700 (33.0)	<0.0001

*Categories B–D represent differences in the availability of human and animal data on the safety of these compounds, these categories do not indicate any difference in drug safety (see Supplemental Table 4 for category definitions).

†Percentages add up to greater than 100% as some patients were treated with multiple drugs.

‡Conditional logistic regression was used to calculate p-values.

FDA, Food and Drug Administration.

Discussion

In this large database analysis of over 10,000 pregnant women, pregnancy in women with lupus was associated with increased resource use and healthcare costs compared to pregnant women without lupus, and also non-pregnant women with lupus. This finding is to be expected, given that women with lupus had a higher risk of pregnancy complications affecting care of the mother, and a doubling of the risk of hypertension, pre-eclampsia, pre-term delivery and stillbirths compared to pregnant women without lupus in this study. This finding is also in line with similar observations of increased complications and higher costs associated with pregnancy in other diseases, such as diabetes19. Thus, additional monitoring and clinical care of pregnant women with lupus is needed in order to appropriately manage the increased risk of complications in these women. However, increased risk of earlier adverse pregnancy events, including miscarriage prior to 22 weeks and threatened miscarriage, appeared modest.

The higher risk of complications associated with pregnant women with lupus highlights the need for regular clinician–patient interaction. This is supported by data from this study showing an increase in reported visits to GPs/pHCPs and OBGYN physicians by pregnant women with lupus. However, surprisingly there was a decrease in rheumatology visits. This decrease may be of concern and could indicate that patients are ignoring their disease whilst focusing on pregnancy. This highlights the need for team-based co-ordination care between OBGYN physicians and rheumatologists. The decrease in rheumatologist visits could also be due to accessibility and affordability reasons. For example, it may be more convenient or cheaper for pregnant women with lupus to see their OBGYN physician rather than their rheumatologist. The increased OBGYN visits (and therefore increased costs) may lead to cuts being made elsewhere, i.e., rheumatology visits. Furthermore, the decrease in rheumatologist visits may be representative of the natural disease course in some patients, where disease activity had decreased, allowing them to consider becoming pregnant and also requiring fewer rheumatologist visits.

The fact that drug use, with the exception of NSAIDs, is consistently lower in pregnant women with lupus compared to non-pregnant women with lupus, without the proportion of women on safer medications increasing, suggests that fewer women with severe active lupus are becoming pregnant.

This study has a number of limitations, mainly due to limitations of the database used in these analyses. The administrative claims database contains information from codes entered by physicians during patient visits, and so does not detect inpatient medications and may be affected by the under-reporting of procedures not affecting reimbursement20. The database did not contain any information on disease activity or severity and, thus, it was not possible to include these factors in the analyses, despite the fact that lupus activity has been linked to poor pregnancy outcomes4–6. Another limitation, given the wide range of organs potentially affected by lupus, was that the database did not capture any information on specific organ manifestations of lupus that occurred over time. This could have provided an important insight into the organ systems most affected during pregnancy. Finally, although we compare pregnant women with lupus to pregnant women without lupus, it should be noted that this control population is not necessarily free from ill health. As there were no disease exclusion criteria, other than a lack of lupus, it is possible that some patients in the control group may have been suffering from other conditions that could have affected their pregnancy.

Conclusions

Our large database analyses have revealed that pregnant women with lupus are at a higher risk of complications, had higher healthcare costs, and were given fewer immunosuppressant medications than the control groups. The increased risk of complications suggests that patients need more guidance from their physicians to ensure that they have the best chance of a successful pregnancy. It is essential that medication management and disease management decisions occur through a shared decision-making process between all parties, where women are fully aware of the potential risks, but also fully informed of what they can do to minimize these in terms of the management of their disease and the timing of their pregnancy.

Transparency

Declaration of funding

UCB Pharma sponsored the study and the development of the manuscript. In addition to content approval by the authors, UCB signed off on the manuscript following a full review to ensure that the publication did not contain any information which has the potential to damage the intellectual property of UCB.

Declaration of financial/other relationships

MP has received grant/research support from UCB Pharma. PD is an employee of the Lupus Foundation of America, Inc. DP is an employee of UCB Pharma.

Acknowledgments

The authors also acknowledge David Friesen (UCB Pharma) for assistance with statistical analyses, and Costello Medical Consulting for editorial and administrative support.