The impact of sex and gender on aortic events in patients with Marfan syndrome

Abstract

Open in new tabDownload slide

OBJECTIVES

The aim of this study was to explore sex and gender differences regarding aortic events in Marfan patients.

METHODS

We analysed all data from our connective tissue disorder database. Only patients with Marfan syndrome were included. For analysis, patients were divided by sex. Female patients were further divided into 2 subgroups: with versus without children. Aortic events were defined as Stanford type A aortic dissection (TAAD) or type B aortic dissection (TBAD) or any aortic intervention.

RESULTS

A population of 183 Marfan patients was analysed for the purpose of this study. One hundred four (57%) were male and 79 (43%) were female patients. Thirty-seven (47%) of the 79 female patients had at least 1 child. Male patients had a significantly higher probability of experiencing an aortic event (P = 0.015) compared to female patients. However, there was no increased probability for recurrent events in male patients compared to female patients (P = 0.063). Follow-up revealed no sex and gender differences in the occurrence of Stanford TAAD or TBAD between male and female patients (P = 0.324/P = 0.534). While 11% of women with children suffered from peripartum aortic events, 24% experienced Stanford TAAD unrelated to pregnancy.

CONCLUSIONS

Male patients have a higher risk of aortic events than female patients. The majority of women were not aware of their Marfan syndrome diagnosis before conceiving. One out of 10 women suffered from peripartum Stanford TAAD or TBAD. Twice as many female patients with children suffered from aortic dissection unrelated to childbirth. There were no sex and gender differences affecting mortality in Marfan patients.

INTRODUCTION

Marfan syndrome (MFS) is a connective tissue disorder inherited in an autosomal dominant fashion. Up to 25% of MFS patients have a de novo mutation. Prevalence is 1 in 5000 individuals, affecting both sexes equally [1]. The diagnosis of MFS is based on the revised Ghent nosology [2]. Patients with MFS show skeletal, ocular and cardiovascular manifestations. Up to 80% of patients with MFS have cardiovascular involvement, including aortic root disease, aortic dissection and mitral valve prolapse with or without mitral regurgitation. Aortic dissection is the main source of morbidity and mortality in patients with MFS.

Data regarding sex and gender-related differences in MFS patients are scarce. Most studies regarding pregnancy in MFS patients do not differentiate between female patients becoming pregnant while being aware of the disease and those who are not. It has been shown that female patients with MFS have an increased risk for aortic dissection during the last trimester of pregnancy and in the early postpartum period [3].

Therefore, we aimed at reporting on sex and gender differences in mortality and aortic events, defined as acute aortic dissection (AAD) or any aortic intervention, in MFS patients.

METHODS

Ethics statement

This study was approved by the cantonal ethics committee of the University of Bern (approval no: 2019-01534). Individual informed consent was obtained from the patients. In cases of minors, consent was obtained from the parent or the legal guardian.

Study population

We retrospectively analysed data from our database of patients with connective tissue disorders. This database comprises all patients with connective tissue disorders that were seen in our tertiary-care referral centre at least once between 1995 and end of 2020. All patients fulfilled Ghent criteria at the time of inclusion into the database.

Definition of aortic events

Aortic events were defined as Stanford type A aortic dissection (TAAD), type B aortic dissection (TBAD) or any aortic intervention. Aortic interventions include open or endovascular surgery involving all aortic segments. While the analysis of aortic events censors each patient after the first event, the recurrent event analysis allows to appreciate the differences in the number of interventions that each patient underwent. All aortic events during a patient’s lifetime, from birth to either death or censoring, were considered.

Definition of peripartum period

According to established standards, the peripartum period was defined as pregnancy and 3 months postpartum [4].

Follow-up

Patients are referred to our tertiary-care centre because of clinically suspected MFS, an aortic event or a family history of MFS. Patients were actively follow-up in our MFS clinic at 3, 6 and 12 months after any aortic event (surgery or dissection) and then, depending on the findings, at least once a year. Patients without previous aortic events were followed on an annual basis with complete aortic imaging performed at least every 3 years. Patients were generally evaluated using echocardiography, computed tomography (CT) angiography or magnetic resonance (MR) imaging.

Surgical approach

Since 1995, we have gradually lowered our threshold for elective aortic root surgery in MFS patients from 55 to 50 mm or less. Currently, the threshold is 45–50 mm in patients suitable for valve-sparing aortic root replacement or those who have progressive aortic dilatation >3 mm per year.

Our operative technique has evolved since 1995 from the modified Bentall technique to the valve-sparing procedure, namely the David reimplantation technique. Prophylactic root replacement was offered to women wishing to conceive if aortic root size exceeds 40 mm. Operative strategy in patients with acute TAAD aimed at the exclusion of any entry tear in the aortic root, ascending aorta or aortic arch. Recent years saw an increase in total arch replacements using the frozen-elephant-trunk procedure to facilitate subsequent treatment of the thoraco-abdominal aorta.

Statistical analysis

Statistical analysis was performed using Stata version 17 (StataCorp, College Station, Tx). For analysis, patients were divided by sex. Female patients were further divided into those with children and those without. Two female patients underwent aortic surgery before pregnancy and, thus, experienced their aortic event before childbirth. These patients were therefore included in the subgroup of female patients without children for descriptive as well as Kaplan–Meier analyses (see below). All other women who had children as well as aortic events experienced their aortic events after childbirth and are thus considered as having children at the time of the event in these analyses. Note that the Cox-regression model described below considers children as a time-varying exposure variable, i.e. women are considered as not having children before childbirth and as having children thereafter.

The distribution of continuous data was assessed by Shapiro–Wilk tests, quantile–quantile plots and histograms. Data are presented as mean (standard deviation), median [quartiles] and number with percentage, according to the type and distribution of the data. Times from patients’ birth to (i) first aortic events or (ii) death were separately analysed by the Kaplan–Meier method followed by a log-rank test to compare the event risk for males versus females, as well as between patients with and without children [5]. To address immortal time bias in comparing women with and without children, we used a Cox-regression model with children as a time-dependent variable as detailed above. Hazard ratios (HR) and their 95% confidence intervals (CI) were estimated from the model, and the proportional hazards assumption was tested using scaled Schoenfeld residuals. To address that several patients had >1 aortic event, recurrent aortic events were analysed by an Andersen–Gill model, with cluster-robust standard errors to account for multiple observations per patient and considering death as a competing risk for occurrence of aortic events. This model also considered children as time-dependent variable in the comparison of women with versus without children. Differences between the groups in continuous outcomes other than time-to-event data were assessed with unpaired Student’s t-test, Mann–Whitney U-test or chi-squared tests, as appropriate. All P-values are two-sided and P-values <0.05 were considered statistically significant, except for subgroup analyses of males versus females with children, males versus females without children and females with children versus females without children, for which a Bonferroni-corrected significance threshold of 0.05/3 = 0.017 was used to denote statistical significance.

RESULTS

Among the 183 MFS patients, 104 (57%) were male and 79 (43%) were female patients. Thirty-seven (47%) of the 79 female patients had at least 1 child (Fig. 1).

Seventy-two percentage of all patients underwent genetic testing and 89% of these carried pathogenic FBN1 mutations.

Aortic events

Of the 104 male patients, 30 {29%, mean age at last follow-up 19.8 [standard deviation (SD): 13.2] years, range 3.4–50.1 years} had no aortic events. Seventy-four (71%) male patients had at least 1 aortic event, including 19 TAAD and 17 TBAD (Fig. 1). Male patients had 151 aortic events (median 1 [first and third quartile: 0, 2], range 0–8) in total.

Thirteen [31%, mean age at last follow-up 15.2 (SD: 18.3) years, range 1.0–71.5 years] of the 42 female patients without children had no aortic events. Twenty-nine (69%) female patients without children had at least 1 aortic event including 4 TAAD and 4 TBAD (Fig. 1). Female patients without children had 42 aortic events (median 1 [0, 1], range 0–3) in total.

Among the 37 female patients with children, 6 [16%, mean age at last follow-up 42.7 (SD: 10.6) years, range 30.2–59.5 years] had no aortic events. Thirty-one (84%) female patients with children had at least 1 aortic event, including 12 TAAD and 10 TBAD (Fig. 1). Female patients with children had 71 aortic events (median 1 [1, 3], range 0–7) in total.

During follow-up, there is a significant difference in the occurrence of an initial aortic event between male and female patients (P = 0.015), with a higher hazard rate of the occurrence of aortic events in male patients (HR 1.56, 95% CI 1.09–2.23, Fig. 2A). However, there is no evidence for an increased probability for recurrent aortic events in male patients compared to female patients (SHR 1.33, 95% CI 0.98–1.79, P = 0.063) (Fig. 3A). In the subgroup analysis, there is a significant difference in the occurrence of an initial aortic event between male patients and female patients with children (HR 2.45, 95% CI 1.55–3.89, P < 0.001); however, there is no significantly increased probability for recurrent aortic events (SHR 1.49, 95% CI 1.04–2.15, P = 0.032) (Figs. 2C and 3C). There is no significant difference in the occurrence of an initial aortic event (HR 0.89, 95% CI 0.57–1.39, P = 0.613) nor an increased probability for recurrent aortic events (SHR 1.13, 95% CI 0.77–1.66, P = 0.542) in male patients compared to female patients without children (Figs. 2B and 3B).

Moreover, female patients without children apparently have a higher probability of experiencing an aortic event during follow-up (HR 2.08, 95% CI 1.24–3.49, P = 0.005) compared to female patients with children (Fig. 2D). However, this difference vanishes after accounting for immortal time bias in the Cox-regression model (HR 1.26, 95% CI 0.70–2.26, P = 0.434). Likewise, there is no increased probability for recurrent aortic events in female patients without children compared to female patients with children (HR 1.32, 95% CI 0.83–2.13, P = 0.240) (Fig. 3D).

Stanford type A aortic dissection

Of the 74 male patients who had at least 1 aortic event, 19 had TAAD [mean age at the time of TAAD 37.6 (SD: 10.9) years, range 18.4–50.3 years]. In 18 patients, TAAD occurred as the first aortic event. Only 1 was aware of his diagnosis prior to TAAD (Table 1). Of the 29 female patients without children who had at least 1 aortic event, 4 had TAAD [mean age at the time of TAAD 36.5 (SD: 21.6) years, range 15.3–64.5 years]. In all 4 patients, TAAD occurred as the first aortic event. Only 1 patient was aware of her MFS diagnosis prior to TAAD (Table 1). Among the 31 female patients with children, who had at least 1 aortic event, twelve had TAAD [mean age at the time of TAAD 48.6 (SD: 14.0) years, range 26.9–69.4 years]. TAAD occurred as the first aortic event in 10 cases. Only 3 patients were aware of their diagnosis prior to TAAD (Table 1).

During the long-term follow-up, there is no statistically significant difference in the rate of TAAD between male and female patients (P = 0.324), male and female patients without children (P = 0.322), male and female patients with children (P = 0.537) nor female patients with children and those without (P = 0.569). Male patients were significantly younger at the time of TAAD compared to female patients with children (P = 0.020).

Stanford type B aortic dissection

Of the 74 male patients, who had at least 1 aortic event, 17 had TBAD [mean age at the time of TBAD 39.0 (SD: 10.0) years, range 24.2–57.4 years]. TBAD as the first aortic event occurred in only 3 patients. Two patients were aware of their diagnosis prior to TBAD (Table 1). Of the 29 female patients without children, who had at least 1 aortic event, 4 had TBAD [mean age at the time of TBAD 42.7 (SD: 10.2) years, range 34.4–56.1 years]. TBAD as the first aortic event occurred in 2 cases and only 1 was aware of her diagnosis prior to TBAD (Table 1).

Among the 31 female patients with children who had at least 1 aortic event, 10 had TBAD [mean age at TBAD 48.4 (SD: 12.3) years, range 23.8–62.5 years]. TBAD as the first aortic event occurred in 3 cases. Two patients were aware of their diagnosis prior to TBAD (Table 1).

During follow-up, there was no statistically significant difference in the occurrence of TBAD between male and female patients (P = 0.534), male and female patients without children (P = 0.652), male and female patients with children (P = 0.604) nor female patients with children and those without (P = 0.989). Male patients were significantly younger at the time of TBAD compared to female patients with children (P = 0.041).

Childbirth

Of the 79 female patients, 37 (47%) had at least 1 child (Fig. 1). The mean maternal age at first childbirth was 25.8 (SD: 3.5) years (range 19.8–34.8 years). Patients reported a total of 79 childbirth (range 0–4, 3 sets of twins). Miscarriages and abortions were excluded.

Prophylactic aortic surgery prior to conception

Prophylactic aortic surgery prior to conception was performed in 2 patients. In both patients, valve-sparing root replacement was performed. One female patient had 1 child, the other had 3 children. Follow-up after childbirth was uneventful in both patients.

Peripartum aortic events

Four (11%) of the 37 female patients with children experienced an aortic event during the peripartum period (Table 2). Three patients had TAAD during the third trimester prior to being diagnosed with MFS. One patient already diagnosed with MFS had uncomplicated TBAD in the postpartum period.

Confirmation of MFS diagnosis before aortic events and childbirth

Among 74 male patients with at least 1 aortic event, 55% (n = 41) were aware of their MFS diagnosis before the first aortic event (Table 1). Among 29 female patients without children with at least 1 aortic event, 79% (n = 23) were aware of their MFS diagnosis before the first aortic event (Table 1). Among 31 female patients with children with at least 1 aortic event, 55% (n = 17) were aware of their diagnosis before the first aortic event (Table 1).

Among 37 female patients with children, only 22% (n = 8) were aware of their MFS diagnosis before their first childbirth. Aortic dissections unrelated to pregnancy occurred in 9 (24%) of the 37 female patients with children, 5 of whom were aware of their diagnosis prior to aortic dissection. The mean time period between last childbirth and TAAD was 27.0 (SD: 12.7) years. The mean time period between last childbirth and TBAD was 24.1 (SD: 10.3) years (Fig. 4).

Follow-up and mortality

The mean follow-up from the first contact until either death or censoring was 3.7 [0.5, 7.7] years for male patients and 4.9 [0.0, 11.5] years for female patients (P = 0.662). The Clark index of all patients was 0.915.

Overall mortality did not significantly differ between male and female patients (HR 0.84, 95% CI 0.30–2.38, P = 0.744), male and female patients without children (HR 0.76, 95% CI 0.18–3.13, P = 0.703), male and female patients with children (HR 0.86, 95% CI 0.28–2.64, P = 0.789) nor female patients without children and female patients with children (HR1.55, 95% CI 0.36–6.63, P = 0.549) (Fig. 5). Thirty-day mortality was 5% (n = 7). The majority of deaths (13 of 16, 81%) occurred in patients with a history of dissection. All causes of death are reported in Supplementary Material, Table S1.

DISCUSSION

The impact of sex and gender on aortic events in patients with MFS has only been reported in a limited number of clinical studies [6–9]. For decades, case series of peripartum AAD or rupture have fostered the notion that women with MFS in general are at a higher risk for aortic events than men. It is only recently that data emerges contradicting this paradigm. Détaint et al. [7] recently reported a large series of MFS patients, where male patients <30 years of age were at higher risk for aortic dilatation and aortic events. Results from the Dutch Concor study [8] and from a nationwide Danish cohort [9] reported additional evidence for male sex and gender being a risk factor in patients with MFS. Recent data from a mouse model for MFS confirm these finding and provide mechanistic insights on a molecular level [10]. Further studies that investigated genotype/phenotype correlations [11–14] showed that, independently of the variants, male patients had more aortic events than female patients.

In our patient population, there were no statistically significant differences in the rate of TAAD or TBAD between male and female patients, similar to the GenTAC registry report [15]. While our data on the rate of TAAD in MFS patients are in line with previous research that reported 16–28% of patients presenting with TAAD [16–18], we are able to provide more granular data regarding sex, gender and childbearing. In our patient population, 18% of male patients, 10% of female patients without children and 32% of female patients with children presented with TAAD. The rate of TBAD, 16% in men, 10% in female patients without children and 27% in female with children in this study is higher than previously reported [6, 19]. Seventy-five percentage of all aortic dissection (TAAD/TBAD) occurred as the first aortic event in patients unaware of the MFS diagnosis at the time of their event. This is higher than previously reported in the Danish cohort [9].

There are little data regarding pregnancy in MFS patients. Only very few reports differentiate between female patients becoming pregnant while being aware of the disease and those who are not. The rate of women becoming pregnant not being aware of MFS diagnosis in this study (78%) is higher than in the GenTAC registry [4] where only 58% of female patients were not aware of their MFS diagnosis before pregnancy. The 2011 European Society of Cardiology guidelines on the management of cardiovascular disease during pregnancy recommend preventive aortic surgery in females with MFS contemplating pregnancy with an aortic root dilatation >45 mm [20]. In the revision of 2018 [3], pregnancy is not recommended in female MFS patients with an aortic dilatation >45 mm. According to the 2010 American Heart Association guidelines on thoracic aortic disease, preventive aortic surgery is recommended in females with MFS who are considering pregnancy with an aortic root dilatation >40 mm [21]. In this study, only 2 patients underwent pre-conception aortic surgery. In both cases, a valve-sparing aortic root replacement using the reimplantation technique was performed. AAD in MFS patients most frequently occurs during the last trimester or early postpartum period, up until 3 months postpartum. However, female MFS patients have an increased risk of dissection after delivery for at least 6 months to 1 year [22]. In the current patient population, 11% of female MFS patients had peripartum events. However, while pregnancy is regarded as an important risk factor for AAD in women with MFS, in our patient population more than twice as many (24% vs 11%) women with children suffered from AADs unrelated to pregnancy. The mean time period between last childbirth was 27.0 (SD: 12.7) years for TAAD and 24.1 (SD: 10.3) years for TBAD. In the current study, there is no evidence for a difference in the number of events—neither primary nor recurrent aortic events—when appropriately modelling child as a time-varying variable.

In patients with a confirmed diagnosis of MFS, pregnancy and childbirth will most probably be uneventful given adequate surveillance is ensured and prophylactic measures according to current guidelines are performed. Nevertheless, women should be counselled that a small but significant risk for peripartum dissection remains. Especially peripartum TBAD is difficult to prevent as most patients dissect at diameters that are far from any threshold for intervention [18]. Factors influencing peripartum risk for aortic dissection have not been fully elucidated. Breastfeeding and its associated rise in oxytocin levels have been identified as a risk factor for aortic dissection in a mouse model for MFS [23], but robust data in humans are lacking.

This study covers a period of >25 years. Awareness for MFS and related disorders has significantly increased over time. Furthermore, growing experience with valve-sparing root replacement has lowered the threshold for prophylactic aortic repair prior to conception. This has also certainly influenced the number of women now becoming pregnant despite the diagnosis of MFS. In counselling patients, the pregnancy-associated risk in women with MFS has to be balanced against the lifetime risk of aortic events in people with MFS in general. The current data show that AAD occurs more frequently unrelated to pregnancy and that the lifetime risk for aortic events in women with children is lower than in men. Identifying women in childbearing age at risk for AAD remains the most important measure to lower mortality in this patient population.

Limitations

The current study carries all limitations associated with the retrospective and observational nature of this study, including the possibility of bias. In particular, confounding cannot be excluded. For example, female MFS patients with a strong phenotype or female MFS patients, who already had aortic events in younger age, may be more reluctant to have children. This is a clear confounder of our analysis.

We looked at a subset of patients who had been carefully chosen for the firm diagnosis of MFS. Although the study as a whole is underpowered to detect differences in survival, the fact that we were able to report a complete follow-up with a mean follow-up of more than a decade suggests that this analysis, despite its exploratory nature, provides important data to better counsel patients and help clinicians make an informed decision.

CONCLUSION

In conclusion, our data suggest that male patients with MFS have a higher risk for aortic events than female patients. Furthermore, 78% of women were not aware of their MFS diagnosis before childbirth. One out of 10 of these women suffered from peripartum AAD. Twice as many female patients with children suffered from AAD unrelated to childbirth. There was no evidence that sex and gender differences affect mortality in MFS patients.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Funding

No funding was acquired.

Conflict of interest: none declared.

Data Availability Statement

Raw data were generated at Inselspital, University of Bern. All relevant data are within the manuscript. Derived data supporting the findings of this study are available from the corresponding author on request.

Author contributions

Maria Nucera: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Visualization; Writing—original draft; Writing—review & editing. Paul P. Heinisch: Conceptualization; Investigation; Methodology; Visualization; Writing—original draft; Writing—review & editing. Bettina Langhammer: Conceptualization; Investigation; Methodology; Writing—review& editing. Silvan Jungi: Visualization; Writing—review & editing. Maks Mihalj: Data curation; Visualization; Writing—review & editing. Patrick Schober: Formal analysis; Visualization; Writing—original draft; Writing—review & editing. Markus M. Luedi: Visualization; Writing—review & editing. Murat Yildiz: Formal analysis; Investigation; Visualization; Writing—review & editing. Florian S. Schoenhoff: Conceptualization; Formal analysis; Methodology; Project administration; Resources; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Joseph S. Coselli, Duke Edward Cameron and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented on 10/09/2020 at the 34th EACTS Annual Meeting.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AAD

  Acute aortic dissection

 
• MFS

  Marfan syndrome

 
• SD

  Standard deviation

 
• SHR

  Sub hazard ratio

 
• TAAD

  Type A aortic dissection

 
• TBAD

  Type B aortic dissection