Marfan syndrome (MFS) is an autosomal dominant connective tissue disorder caused by a mutation in the FBN-1 gene encoding for the glycoprotein, fibrillin (1). While MFS is associated with diverse clinical manifestations in multiple organ systems, it is the cardiovascular complications of aortic dilation and dissection that account for the morbidity and mortality associated with the disease (2). Risk factors for adverse aortic outcomes, namely the need for aortic surgery, aortic insufficiency, and aortic dissection, have been identified and include greater aortic size (2), obesity (3), family history of aortic dissection (2), and older age (2). The impact of gender on risk remains controversial with some studies noting female gender, and others, male gender, as a risk factor for aortic dissection.

Numerous case reports and 8 retrospective studies have documented aortic dissection during, or immediately following, pregnancy in women with MFS (4- 9). In the majority of these women the diagnosis of MFS was unknown at the time of pregnancy (4- 9). As a result of this retrospective data, which are inherently subject to selection bias, women with MFS in the United States are counseled to avoid pregnancy, or alternatively, undergo surgical ascending aortic replacement prior to conception, if the aorta measures >4 cm (10). Perhaps as a result, there is very little prospective data assessing the impact of pregnancy on aortic growth and clinical outcome in women with MFS. We sought to assess the impact of pregnancy on the rate of aortic growth as well as on short- and long-term clinical outcomes in a group of women with MFS followed by a cardiologist prospectively throughout pregnancy.

An institutional review board–approved retrospective review of clinical records on all patients seen in a tertiary cardiology subspecialty aortopathy clinic over the past 10 years was performed. The aortopathy clinic is a regional referral center for a 6-state region. Obstetrical and cardiac data are stored within a statewide electronic medical record system.

Detailed echocardiographic and demographic data on all patients satisfying inclusion criteria were obtained and included for review. For inclusion into the study the patient needed to fulfill Ghent criteria for a diagnosis of MFS either with or without confirmatory genetic testing (11), be a female of ≥18 years of age having attained final height as evidenced by no change in height over the year prior to initiation of data collection, and have had a clinic visit within the past 18 months or documentation of death. Women with a documented TGFBRI or TGFBRII mutation, or clinical features suggestive of such in the absence of genetic testing, including arterial tortuosity, suggestive facial features, or abnormal uvula (12), were excluded from data collection.

Subjects were categorized as those having had, and those not having had, a previously documented pregnancy irrespective of outcome. Women with a prior pregnancy were further subdivided into those receiving, and those not receiving, prospective cardiac care during pregnancy. Women followed prospectively with a confirmed diagnosis of MFS prior to pregnancy, underwent echocardiography at a minimum of once within the 12 months prior to pregnancy, at 12 to 16 weeks' gestation, 24 to 26 weeks' gestation, 35 to 39 weeks' gestation, 4 to 16 weeks postpartum, and then annually as is our clinical protocol. Additional echocardiograms were performed if there was clinical cause for concern. Nulliparous females and females not receiving prospective cardiac care during pregnancy had a minimum of annual echocardiograms from the time of diagnosis until last follow-up.

Data on 2 clinical outcomes was recorded: 1) a composite adverse outcome, defined as death, aortic dissection, severe symptomatic aortic insufficiency, or need for urgent surgery; and 2) the need for elective aortic surgery during the period of clinical follow-up.

For all 3 patient groups: 1) pregnancy followed prospectively; 2) pregnancy not followed prospectively; 3) no prior pregnancy, the following data were recorded: age at diagnosis, type and duration of medical therapy, presence or absence of a family history of MFS, and duration of follow-up after first pregnancy, or in the case of nulliparous controls, after age 18 years.

Cardiac and obstetric data collection varied dependent on patient group but in all cases was acquired prospectively and documented in the patient's obstetrical or cardiac medical records. For pregnant females receiving prospective cardiac care, cardiac data recorded included age, blood pressure, weight, height, body surface area, presence or absence of cardiac symptoms, aortic root diameter, presence or absence of aortic dissection, and degree of aortic insufficiency at: 1) time of initial visit at age ≥18 years with final height attained; 2) subsequent visits prior to pregnancy; 3) all visits during pregnancy as defined previously; 4) visit 4 to 12 weeks postpartum; 5) annual visits subsequent to a pregnancy; and 6) last follow-up visit prior to surgery or death.

For pregnant females not receiving prospective cardiac care, recorded data included age at each pregnancy, current age, blood pressure at peripartum visits, most recent blood pressure, current weight, height, body surface area, presence or absence of aortic dissection, or need for elective aortic replacement. Aortic diameter prior to any pregnancy (when available) and most recent aortic diameter prior to surgery or adverse outcome were obtained.

For all females with a prior pregnancy, cardiac and obstetrical records were reviewed to determine the interval between pregnancies, the presence of maternal pregnancy complications including fetal loss, ectopic pregnancy, gestational diabetes, pregnancy-induced hypertension, pre-term labor <37 weeks, and the need and indication for cesarean section. Neonatal data including gestational age at delivery, birth weight, and presence or absence of a subsequent diagnosis of Marfan syndrome in the offspring was recorded.

A control group of adult women who had not had a prior pregnancy was identified. Demographic data recorded included age, blood pressure, weight, height, body surface area, presence or absence of cardiac symptoms, aortic root diameter and degree of aortic insufficiency at initial adult visit as defined previously, and at last follow-up visit prior to surgery or death.

Echocardiographic review included documentation of the sinus of Valsalva diameter in systole measured off-line from parasternal long-axis 2-D images using inner-edge-to-inner-edge technique (13). Aortic insufficiency was noted to be absent, trivial, mild, moderate, or severe based on aortic insufficiency jet to annulus ratio. Aortic diameters were expressed as absolute values, as indexed values based on age and body surface area, and as z-scores based on normative data for body surface area where appropriate. Indexed values were calculated in order to control for the normal increase in aortic size with age, and for increase in somatic size over time when in the nonpregnant state (14). Changes in absolute aortic diameter over time were calculated as the change in absolute aortic diameter divided by days between echocardiographic studies and expressed as millimeters per month.

All women followed prospectively were followed by cardiology and maternal fetal medicine throughout the pregnancy. Women were seen on a monthly basis by cardiology with echocardiographic frequency as noted previously. Metoprolol 25 mg twice daily (bid) was recommended for prophylaxis against progressive aortic dilation to all women irrespective of baseline blood pressure. Medication was uptitrated if required to maintain a systolic blood pressure <120 mm Hg. Hypertension was defined as a systolic or diastolic blood pressure >135 mm Hg or 85 mm Hg, respectively. All women with elevated blood pressure refractory to medical management were placed on in-hospital bed rest. All women with significant aortic dilation, defined as ≥5.0-mm increase in maximal aortic diameter from baseline with attainment of an absolute aortic diameter >4 cm, were delivered if fetal viability was expected. Cesarean section or a passive second stage of labor with regional anesthesia (absence of pushing, with forceps or vacuum extraction) at term (>37 weeks) was otherwise recommended. Induction of labor was avoided unless there was obstetrical reason to proceed with such. Peripartum recommendations included hospital admission until a minimum of 72 h post-delivery and an echocardiogram prior to discharge. All women were seen in a subspecialty cardiology clinic within 2 months following delivery. Women receiving metoprolol were maintained on therapy for 3 months following delivery with the dose adjusted to maintain a blood pressure of <120 mm Hg without symptoms of hypotension.

Analysis was conducted using Stata 10.0 (Stata Corp., College Station, Texas). Continuous data are expressed as mean ± SD or median (interquartile range) as appropriate and categorical data were tabulated. A 2-sided p value <0.05 was considered statistically significant. Demographic and echocardiographic data for women with prior pregnancy versus nulliparous females was compared using chi-square test for dichotomous or categorical variables and t test or rank sum test for continuous variables depending on their normality. Echocardiographic measurements used for these initial analyses included aortic root diameter at initial echocardiogram in adulthood and aortic root diameter at last echocardiogram during follow-up or prior to surgical intervention or death. The primary outcome measures were measured as dichotomous variables and were compared for the 2 groups using chi-square analysis. The association of clinical factors with adverse cardiac outcome and need for elective surgery was analyzed by logistic regression. Factors that had a p value <0.2 were included in a multiple logistic regression model to determine the adjusted odds ratio. Variables were retained if they changed the point estimate by more than 10%. Echocardiographic and demographic variables associated with maternal or fetal pregnancy complications were also examined using chi-square, t, or rank sum test as appropriate.

For the women followed prospectively during pregnancy, differences in aortic growth rates were calculated for the following time periods: 1) baseline: last aortic root measurement prior to pregnancy less initial adult measurement divided by the intervening time interval; 2) pregnancy: last recorded aortic root measurement in pregnancy or immediately post-pregnancy (whichever was larger) less initial pregnancy measurement divided by the intervening time interval; 3) post-pregnancy: last recorded aortic root measurement at follow-up or prior to surgery or dissection less initial postpartum measurement divided by the intervening time interval.

Aortic growth rates for the 3 time periods were analyzed for equality of variances and then compared using both a paired t test with unequal variance and a Pearson correlation coefficient. Aortic growth rates between women followed prospectively during pregnancy and nulliparous women were compared.

Of the 98 women who were followed in the aortopathy clinic, and whom satisfied the inclusion criteria outlined previously, 69 (72%) experienced a total of 199 pregnancies resulting in 170 (86%) live births, 26 (13%) spontaneous abortions, and 2 ectopic pregnancies. There were no elective terminations, which is reflective of the low overall pregnancy termination rate of the state. The pregnancy cohort experienced between 1 and 12 pregnancies (median 3). The remaining 29 women did not have a prior pregnancy and served as the nulliparous control group. No woman was lost to follow-up care.

Of those women with a prior pregnancy, 35 (51%) were followed prospectively during the pregnancy with routine echocardiography whereas 34 (49%) women had prior pregnancies without routine peripartum cardiac evaluation. Of those not routinely followed by a cardiologist during pregnancy, 30 (88%) did not receive such care as the diagnosis of maternal Marfan syndrome was unsuspected at the time of pregnancy.

Of the 170 pregnancies resulting in live births, obstetrical complications occurred in 17 (10%) including pregnancy-induced hypertension in 9 (5%), hyperemesis in 1 (<1%), and premature labor in 7 (4%). Cesarean section was performed in 20 (12%). In all cases, this was a planned procedure for concern over maternal health secondary to aortic dilation >4 cm, or repeat procedure. All other women receiving routine peripartum cardiac care underwent a vaginal delivery with vacuum or forceps extraction. There was no relationship between the presence of obstetrical complications and maternal age (21.5 vs. 25.4 years, p = 0.8), interval between pregnancies (15.0 vs. 15.5 months, p = 0.5), or pre-pregnancy aortic diameter (36.9 vs. 35.4 mm, p = 0.33). The frequency of pregnancy complications did not differ from that of the general population of the state (15). Neonatal outcomes included small for gestational age in 9 (5%), prematurity <37 weeks in 10 (6%), neonatal death secondary to complications of extreme prematurity in 2 (1%), and congenital cystadenomatoid malformation in 1 (<1%). Adverse fetal outcome was associated with a higher rate of beta-blocker usage (28% vs. 7%, p = 0.002), although women with a larger aorta during pregnancy were more likely to be receiving beta-blocker medicine (p = 0.0004). Maternal age at delivery and interval between pregnancies were not associated with adverse fetal outcome. Of the 167 offspring surviving the neonatal period, 125 (75%) underwent clinical and echocardiographic evaluation for MFS. A confirmatory diagnosis was made in 88 (70%) all of whom had cardiac involvement.

Within the entire cohort of women experiencing a pregnancy there were no acute aortic dissections. Two women, both of whom were known to have MFS but were not receiving routine cardiac care throughout pregnancy, developed symptomatic carotid artery dissections; 1 presented with symptoms of a transient ischemic attack in the third trimester, and the other an enlarging pulsatile neck mass after routine vaginal delivery without assistance. No woman required intrapartum cardiovascular surgery or pregnancy termination. One patient with a 49-mm aorta and mild aortic insufficiency entering into the pregnancy developed New York Heart Association functional class II symptoms in the face of severe aortic insufficiency at 38 weeks' gestation and went on to have aortic root replacement at 6 months postpartum. Of the 10 pregnancies complicated by premature delivery, early delivery occurred as a result of pre-term labor in 6, pregnancy-induced hypertension in 2, and progressive aortic dilation (>5 mm from baseline) with an absolute aortic diameter of >4.0 cm in 2. Both of the latter 2 women were delivered at 36 weeks' gestation.

Complete serial echocardiography with a minimum of 3 pre-pregnancy echocardiograms, 3 intrapartum echocardiograms, and 3 post-partum echocardiograms was available for 55 pregnancies in 35 women. Three women each experienced 1 pregnancy following prior elective ascending aortic replacement and were thus excluded from analyses assessing aortic growth rate. Two of the 3 had a valve-sparing operation while the other had a composite graft with a mechanical valve. Pregnancy was complicated by progressive neoaortic insufficiency from trivial to moderate in both women with valve-sparing procedures.

Excluding the patients with prior aortic surgery, baseline aortic root diameter within the year prior to pregnancy was 36.1 ± 4.4 mm with 14 (27%) of the pregnancies entered into with an aortic root diameter of ≥4.0 cm. Aortic diameter increased by 3 mm (interquartile range: 0 to 7 mm) during the course of the pregnancy resulting in an aortic root diameter of 39.0 ± 6.0 mm at the time of delivery. Aortic growth rate increased significantly during the intrapartum period and then fell following delivery, albeit remaining significantly above the baseline rate (Figure 1).

As noted in (Table 1), women with and without a prior pregnancy were well matched with respect to initial age, body size, and aortic size. Women who were not pregnant, however, were more likely to be compliant with recommendations for medical therapy to prevent aortic dilation. At the time of last follow-up, women who had a prior pregnancy had a significantly greater aortic diameter (Table 1).

Despite the lack of catastrophic peripartum complications, the prevalence of both adverse outcome and elective aortic surgery during long-term follow-up was higher in those women who had a prior pregnancy (Table 2). This finding is in keeping with the higher rate of aortic growth documented during pregnancy compared with each woman's prior baseline aortic growth rate (Figure 1).

Factors associated with long-term adverse cardiovascular outcome on univariate analysis within the entire cohort of patients with prior pregnancy included greater absolute and indexed aortic size at onset of pregnancy, initial pregnancy aortic diameter >4 cm, greater rate of aortic growth, increased number of pregnancies, lack of beta-blocker use during pregnancy, and lack of prospective pregnancy follow-up (Table 3). On multivariate analysis, initial aortic diameter and rate of change in aortic diameter (log) remained the only independent predictors of long-term adverse cardiovascular outcome (Table 4).

The data on women with MFS receiving cardiac care during pregnancy is extremely limited. In the 2 studies to date, no woman under the care of a cardiologist experienced dissection of the ascending aorta (15- 16). These authors concluded that pregnancy was safe and did not typically impose a deleterious health impact in unoperated women with MFS with initial aortic diameters of ≤45 mm (16). Our study confirms the findings of the 2 prior prospective studies in that no woman with a known diagnosis of MFS, who was under the care of a cardiologist, experienced acute aortic dissection during the 199 pregnancies followed. While there are obvious potential risks associated with pregnancy in these women, these risks may be minimized by timely diagnosis prior to conception, tight blood pressure control, use of beta-blockade despite normotension, close echocardiographic surveillance, and delivery strategies aimed at reducing aortic wall stress.

Despite the absence of catastrophic peripartum complications, we did note an increase in long-term adverse outcomes in those women with MFS who underwent pregnancy. The risk was greatest in those with multiple pregnancies, those not receiving beta-blocker therapy during pregnancy, and those with a larger aortic dimension at baseline. A larger increase in aortic diameter during pregnancy heralded an increase in long-term adverse outcome. The increase in aortic size during pregnancy is not unique to women with MFS, but is known to occur during normal healthy pregnancies (17) and with increased magnitude in women with pre-eclampsia (17). In contrast to the woman without MFS wherein increases in aortic diameter of up to 1 mm have been reported (17), our patients had greater aortic diameters at baseline with up to a 7-mm increase in aortic root size during pregnancy.

Despite the lack of data supporting the often quoted “10% incidence of aortic dissection during pregnancy if the aorta is ≥4 cm” (6), and documentation of distal aortic dissection in pregnant women with MFS following aortic root replacement (18- 19), the 2010 thoracic aortic guidelines advocate avoidance of pregnancy if the aortic root is >4.0 cm and recommend surgical ascending aortic replacement in those who desire pregnancy (10). These recommendations are on the one hand more stringent, and on the other more liberal, than the European (20) and Canadian (21) guidelines, which report an aortic diameter of <45 mm to be considered safe for pregnancy, and make specific mention of the ongoing risk of aortic dissection in the pregnant patient following aortic root replacement. In the 2 prospective studies to date (15- 16), the only women to experience aortic dissection during pregnancy were 2 who had prior aortic root replacement. At present there is no data to support a recommendation for early prophylactic surgery solely for the purposes of child-bearing. In addition, the feasibility of monitoring for distal disease in the pregnant women following ascending aortic replacement, is fraught with difficulty. Routine computed tomography angiography is contraindicated, and serial routine magnetic resonance imaging during pregnancy is both costly and difficult for the pregnant woman.

The current literature assessing aortic growth rate during pregnancy in women with MFS is limited to 10 women noting no difference in aortic growth rate immediately prior to, or following pregnancy (16). In contrast, our data demonstrate a consistent increase in aortic growth rate during pregnancy in women with MFS. In contrast to the prior studies, our patient cohort had serial echocardiographic monitoring throughout pregnancy, experienced a greater number of pregnancies per woman, and were followed over a longer period of time. Despite this, however, patient numbers were limited and thus the statistical model used may be limited to overfitting.

In light of the data presented herein, we counsel our patients with MFS with regard to both the short-term and long-term risks of pregnancy. Women are advised that their aorta will most likely increase in size during pregnancy and is not likely to return to baseline size. Subsequent pregnancies will further the degree of aortic dilation. Given that the larger aorta dilates at a more rapid rate, women can anticipate a higher likelihood of experiencing an adverse cardiovascular outcome in the long term. The risk of death or acute ascending aortic dissection during pregnancy or the postpartum period is extremely low in women who have not undergone prior ascending aortic root replacement and in whom the aortic diameter is <45 mm. Careful monitoring and multidisciplinary management during pregnancy is beneficial.