Marfan syndrome is associated with increased risk for gynecologic disorders and maternal complications

Abstract

Purpose

To determine whether patients with Marfan syndrome are at an increased risk for reproductive disorders.

Methods

This retrospective cohort study was conducted using the US collaborative network on the TriNetX research network of health care organizations. We included female patients aged 18–44 and identified a cohort of 4347 patients with Marfan syndrome (ICD-10 Q87.4). Our control cohort consisted of 16,424,990 patients without a diagnosis of Marfan syndrome or Ehlers-Danlos syndrome (ICD-10 Q79.6). The primary outcomes included gynecologic diagnoses such as dysmenorrhea and endometriosis, and our secondary outcomes included urogynecologic, fertility, and obstetric outcomes, all identified by ICD-10 codes. We conducted a relative risk analysis with a p-value of <0.05 considered significant.

Results

Patients with Marfan syndrome were at an increased risk for pelvic and perineal pain, dysmenorrhea, abnormal uterine bleeding, endometriosis (all p <0.0001), dyspareunia (p =0.0009), leiomyoma (p =0.0076), polyp of female genital tract (p =0.016), urinary incontinence (p <0.0001), female genital prolapse (p =0.0006), fertility testing (p =0.0075), cesarean delivery (p =0.0003), gestational hypertension (p =0.0012), and pre-eclampsia (p =0.0024) compared to the control group following an adjusted, matched comparison.

Conclusions

Patients with Marfan syndrome have an increased risk of numerous reproductive disorders and obstetric complications compared to patients without this diagnosis.

Introduction

Marfan syndrome (MFS) is an inherited connective tissue disorder caused by mutations in the gene for fibrillin-1, an extracellular matrix protein, which plays a critical role in the stabilization of connective tissue through the formation of microfibrils in addition to regulating the cytokine bioavailability of transforming growth factor-B (TGF-β) [1]. The disease is inherited in an autosomal-dominant pattern and is estimated to affect 1 in every 4000–6000 individuals [2]. The commonly documented clinical manifestations of MFS include instability of the aortic root, ocular lens dislocation, arachnodactyly, and tall stature.

The majority of the literature regarding obstetric and gynecologic manifestations of MFS is focused on serious maternal complications in this population, due to the high risk of aortic aneurysm and rupture during pregnancy and childbirth [3]. Additionally, a connection between MFS and pelvic floor disorders has been reported, with a high proportion of individuals with this disease experiencing symptoms such as urinary incontinence and pelvic organ prolapse [4, 5]. However, the pathophysiology of MFS warrants further exploration into other gynecologic and reproductive manifestations of the disease. This is particularly true for the role of fibrillin-1 in regulating TGF-β. The increased deposition and signaling of TGF-β have been suggested to underpin several manifestations of MFS, such as weakness of the aortic wall and tall stature [6]. This cytokine has also been suggested to play a role in the normal physiology of the female reproductive system, including endometrial repair following menses and follicular maturation and ovulation, and its dysfunction has been suggested to contribute to several gynecologic disorders, including endometriosis, fibroids, and heavy menstrual bleeding [7–9]. Despite these potential connections, investigations into gynecologic manifestations of MFS are limited, due in part to the rarity of the condition.

The primary objective of this investigation was to utilize a large database to evaluate gynecologic diagnoses in patients with MFS. Given the role that TGF-β plays in the female reproductive system and in gynecologic diseases, we hypothesized that the risk of gynecologic diagnoses would be higher among individuals with MFS compared to the general population. The secondary objectives of the study were to evaluate the risk of urogynecologic, fertility, and obstetric diagnoses in MFS patients, looking to support findings documented in previous literature with smaller sample sizes.

Materials and methods

Database description

We utilized the TriNetX research network to conduct this retrospective cohort study, accessing the database from February 1–9, 2024, to May 22, 2024. We used the USA Collaborative Network within the TriNetX platform to establish our patient cohorts. This network contains deidentified electronic health record (EHR) data from over 100 million patients from 64 US health care organizations (HCOs). HCOs join the network to receive access to the platform’s analytics, deidentified data, and opportunities for sponsored clinical trials [10]. Data is deidentified per the Health Insurance Portability and Accountability Act (HIPAA) criteria—Section §164.514(a) of the HIPAA Privacy Rule. TriNetX has an exemption from the Western Institutional Review Board due to the data being deidentified, complying with HIPAA, and hundreds of peer-reviewed studies have been published based on data from the network [10].

Study population

The total study population consisted of 18–44-year-old female patients who received care in the participating US health care systems in the previous 10 years. Of note, ICD-10 codes replaced ICD-9 codes in October 2015, so the study population ultimately includes patients since the adoption of ICD-10 coding. Inclusion into the MFS cohort was defined as a documented diagnosis of Marfan syndrome per ICD-10 Q87.4, and patients could have been seen in any capacity in the health care system, including both inpatient and outpatient encounters. The control cohort was defined as patients who did not have a diagnosis of Marfan syndrome (ICD-10 Q87.4) or Ehlers-Danlos syndromes (ICD-10 Q79.6) due to confounding disease presentations. These cohorts were created in February 2024. For the MFS and control obstetric cohorts, an additional inclusion criterion was defined to only include patients who have been pregnant by either ICD-10 10 (pregnancy) or Z33 (pregnant state). Cohorts to compare diagnoses within MFS syndrome patients by race and ethnicity were also created for White, Black, Asian, and Hispanic patients (Fig. 1). These additional cohorts were created in May 2024. We did not create cohorts for American Indian/Alaskan Native and Native Hawaiian due to limited patient numbers in the MFS cohort and for other and unknown races due to limitations in data interpretation.

Fig. 1.

Flow diagram of race/ethnicity cohorts. Legend: MFS, Marfan syndrome

Covariates and outcomes

MFS and control patients were matched 1:1 for current age for gynecologic analyses and for current age and history of a pregnancy through assisted reproductive technology (ART) for obstetric analyses. The built-in TriNetX propensity score matching function was used which offers matching only in a 1:1 ratio and not 2:1 or 3:1 despite the size of the database. Due to no differences found by race/ethnicity, we did not match for these demographic characteristics. We compared the prevalence of various gynecologic and urogynecologic conditions, outcomes related to infertility/ART, and obstetric outcomes between the cohorts using ICD-10, CPT, and logical Observation Identifiers Names and Codes (LOINC) codes. The included gynecologic conditions were pelvic and perineal pain (ICD-10 R10.2), unspecified dysmenorrhea (N94.6), dyspareunia (N94.1), abnormal uterine and vaginal bleeding (N93), excessive, frequent, and irregular menstruation (N92), unspecified irregular menstruation (N92.6), iron deficiency anemia secondary to chronic blood loss (D50.0), endometriosis (N80), leiomyoma (D25), and polyp of female genital tract (N84). Urogynecologic conditions and procedures included female genital prolapse (N81), urinary incontinence (N39.3 and N39.4), and fecal incontinence (R15). The infertility and ART outcomes included were supervision of a pregnancy resulting from ART (ICD-10 O09.819, O09.81, O09.811–3), encounter for fertility testing (ICD-10 Z31.41), encounter for ART procedure cycle (ICD-10 Z31.83), presence of a result from an anti-Mullerian hormone (AMH) laboratory test (LOINC 38476–8), and oocyte retrieval or embryo transfer (CPT 58970, 58974). The obstetric outcomes assessed were cesarean delivery procedures (CPT 1008991), postpartum hemorrhage (PPH) (ICD-10 O72), spontaneous abortion (ICD-10 O03), gestational hypertension (ICD-10 O13), pre-eclampsia (ICD-10 O14), and preterm prelabor rupture of membranes (PPROM) (ICD-10 O42.0).

Statistical analysis

We compared the risk of the aforementioned outcomes between patients with MFS and the control cohort using z-tests, risk ratios (RRs), and 95% confidence intervals (CIs). Statistical analyses were conducted in the TriNetX platform using R version 4.0.2, Survival package v3.2–3. Z-tests were used to test the null hypothesis that the risk difference of each outcome was equal to 0. p values below 0.05 were considered statistically significant. Sub-analyses included a comparison of diagnoses within the MFS cohort by race/ethnicity.

Results

Gynecologic diagnoses

A total of 16,429,337 patients were included in this study, 4347 in the MFS cohort and 16,424,990 patients in the control cohort. The mean (standard deviation (SD)) current age of the MFS cohort was 29.3 (7.78), and the mean (SD) current age of the control cohort was 31.2 (7.51). Demographics for each cohort prior to matching are seen in Table 1. After matching for age, there were 3712 patients in each cohort, with a mean (SD) age of 29.3 (7.78) for each. Our cohorts stratified by race had 3512 White MFS patients, 549 Black MFS patients, 133 Asian MFS patients, 494 Hispanic MFS patients, and 3686 non-Hispanic MFS patients. Our pregnant MFS cohort had 533 patients, before matching, compared to a pregnant control cohort of 1,752,623 patients.

Table 1.

Baseline characteristics of the study population

	MFS (N = 4347)	Control (N = 16,424,990)
Age, years	29.3 ± 7.8	31.2 ± 7.5
Race
White	2606 (70.2)	8,723,399 (55.9)
Black or African American	384 (10.4)	2,472,715 (15.8)
Asian	94 (2.5)	661,022 (4.2)
American Indian/Alaskan Native	20 (0.5)	75,150 (0.5)
Native Hawaiian or Other Pacific Islander	12 (0.3)	65,225 (0.4)
Other race	183 (4.9)	963,886 (6.2)
Unknown race	413 (11.1)	2,658,448 (17.0)
Ethnicity
Hispanic	335 (9.6)	1,922,908 (12.3)
Not Hispanic	2689 (72.4)	8,771,678 (56.2)
Unknown ethnicity	668 (18.0)	4,925,250 (31.5)

Statistics presented as (mean ± SD) for age and n (section of column %) for race and ethnicity

MFS, Marfan syndrome

The risk of pelvic and perineal pain (RR 1.92 [95% CI (1.61, 2.28)]), unspecified dysmenorrhea (RR 1.73 [95% CI (1.35, 2.23)]), dyspareunia (RR 2.27 [95% CI (1.38, 3.75)]), abnormal uterine bleeding (RR 2.05 [95% CI (1.661, 2.554)]), irregular menstruation (RR 1.62 [95% CI (1.34, 1.97)]), excessive, frequent, and irregular menstruation (RR 1.77 [95% CI (1.52, 2.06)]), and iron deficiency anemia secondary to chronic blood loss (RR 3.81 [95% CI (2.20, 6.60)]) were increased in the MFS cohort compared to the control cohort (Table 2). The risk of endometriosis (RR 2.71 [95% CI (1.65, 4.47)]) was higher in the MFS cohort than in the control cohort. The MFS cohort also had a higher risk of leiomyoma (RR 1.89 [95% CI (1.17, 3.03)]) and polyp of the female genital tract (RR 2.40 [95% CI (1.15, 5.01)]). For the urogynecologic outcomes, MFS patients had an increased risk of female genital prolapse (RR 2.50 [95% CI (1.45, 4.31)]) and urinary incontinence (RR 2.92 [95% CI (1.84, 4.63)]) but not fecal incontinence (RR 2.10 [95% CI (0.99, 4.45)]). The MFS cohort was additionally more likely to have an encounter for fertility testing (RR 2.60 [95% CI (1.26, 5.38)]) and a prior AMH test (RR 2.17 [95% CI (1.10, 4.29)]) but not supervision of an ART pregnancy, encounter for an ART procedure cycle, oocyte retrieval, or embryo transfer.

Table 2.

Risk of gynecology, urogynecology, fertility, and obstetric outcomes in MFS patients versus control

	MFS	Control	RR	95% CI
Gynecology, n = 3712 per cohort
Pelvic and perineal pain	343 (9.4)	179 (4.8)	1.92	(1.61, 2.28)
Dyspareunia	50 (1.4)	22 (0.6)	2.27	(1.38, 3.75)
Abnormal uterine and vaginal bleeding	243 (6.6)	118 (3.2)	2.06	(1.66, 2.55)
Excessive, frequent, and irregular menstruation	432 (11.6)	244 (6.6)	1.77	(1.52, 2.06)
Unspecified dysmenorrhea	163 (4.4)	94 (2.5)	1.73	(1.35, 2.23)
Unspecified irregular menstruation	255 (6.9)	157 (4.2)	1.62	(1.34, 1.97)
Iron deficiency anemia, secondary to chronic blood loss	61 (1.6)	16 (0.4)	3.81	(2.20, 6.60)
Endometriosis	57 (1.5)	21 (0.6)	2.71	(1.65, 4.47)
Leiomyoma	49 (1.3)	26 (0.7)	1.89	(1.17, 3.03)
Polyp of female genital tract	24 (0.7)	10* (0.3)	2.40	(1.15, 5.01)
Urogynecology, n = 3712
Urinary incontinence	70 (1.9)	24 (0.7)	2.92	(1.84, 4.63)
Female genital prolapse	45 (1.2)	18 (0.5)	2.50	(1.45, 4.31)
Fecal incontinence	21 (0.6)	10* (0.3)	2.10	(0.99, 4.45)
Fertility care, n = 3712
Encounter for fertility testing	26 (0.7)	10* (0.3)	2.60	(1.26, 5.38)
Presence of an AMH laboratory test	26 (0.7)	12 (0.3)	2.17	(1.10, 4.29)
Encounter for ART procedure cycle	15 (0.4)	10* (0.3)	1.50	(0.68, 3.34)
Oocyte retrieval or embryo transfer	14 (0.4)	10* (0.3)	1.40	(0.62, 3.15)
Supervision of a pregnancy resulting from ART	17 (0.5)	10* (0.3)	1.70	(0.78, 3.71)
Obstetric outcomes, n = 523
Gestational hypertension	44 (8.4)	19 (3.6)	2.32	(1.37, 3.91)
Pre-eclampsia	42 (8.0)	19 (3.6)	2.21	(1.30, 3.75)
Spontaneous abortion	32 (6.1)	26 (5.0)	1.23	(0.74, 2.04)
PPROM	20 (3.8)	17 (3.3)	1.18	(0.62, 2.22)
Cesarean delivery	85 (16.3)	46 (8.8)	1.85	(1.32, 2.59)
PPH	33 (6.3)	21 (4.0)	1.57	(0.92, 2.68)

n (% of cohort). RR, relative risk; 95% CI, 95% confidence interval

^*Any number of patients ≤ 10 is reported as 10 in TriNetX

All outcomes matched for current age, obstetric outcomes also matched for history of pregnancy resulting from assisted reproductive technology

MFS, Marfan syndrome; AMH, anti-Mullerian hormone; ART, assisted reproductive technology; PPROM, preterm prelabor rupture of membranes; PPH, postpartum hemorrhage

Obstetric outcomes

After matching, there were 523 patients in both the obstetric MFS and control cohorts. Patients in the MFS cohort were at a higher risk for cesarean delivery procedures (RR 1.85 [95% CI (1.32, 2.59)]), gestational hypertension (RR 2.32 [95% CI (1.37, 3.91)]), and pre-eclampsia (RR 2.21 [95% CI (1.30, 3.75)]) (Table 2). However, there was no difference between the risk for PPH, spontaneous abortion, and PPROM between the two cohorts.

Comparison of diagnoses by race and ethnicity

Within the MFS cohort, all the diagnoses were compared in Black versus White, Asian versus White, and non-Hispanic versus Hispanic patients. After matching for age, there were 483 patients in the Black versus White analysis, 111 in the Asian versus White analysis, and 413 patients in the Hispanic versus non-Hispanic analysis. Black MFS patients were more likely than White MFS patients to have unspecified dysmenorrhea (RR 2.50 [95% CI (1.22, 5.14)]) and irregular menstruation (RR 1.63 [95% CI (1.11, 2.39)]). There were no significant differences between Asian and White and Hispanic and non-Hispanic MFS for any of the diagnoses.

Discussion

Our study tested the hypothesis that patients with a diagnosis of MFS would have a higher incidence of gynecologic, urogynecologic, fertility, and obstetric disorders compared to a control cohort [7–9]. Our study utilized a database of deidentified patient information derived from multiple large hospital systems, making this (to our knowledge) the largest analysis of MFS in the context of obstetrics and gynecology to date. Overall, we found a statistically significant increased risk for many life-impacting gynecologic and urogynecologic diagnoses. Individuals with MFS were also more likely to have a history of fertility testing and a recorded AMH lab value than their matched controls, but there was no difference between the two groups regarding encounters for ART procedures, oocyte retrieval or embryo transfer, or supervision of a pregnancy resulting from ART. For obstetric outcomes, there was an increased risk for cesarean delivery, gestational hypertension, and pre-eclampsia in MFS patients, but there was no difference in the risk of spontaneous abortion, PPROM, or PPH.

As there have been well-documented health disparities in the incidence, diagnosis, and treatment of many of these gynecologic and obstetric conditions, we expected to find differences when comparing MFS patients with different racial and ethnic backgrounds. However, our secondary analysis of racial and ethnic differences within the MFS cohort found very few differences. This may be due to the small sample size of only a few hundred patients in each comparison.

In our study, we found an increased risk for specific benign gynecologic diagnoses in MFS patients, associations which to our knowledge have not been reported in the literature, including menstrual cycle irregularities, endometriosis, leiomyoma, and uterine polyps. Interestingly, these conditions have previously been implicated in dysfunctional TGF-β signaling, in line with our proposed hypothesis. The fact that this has not been observed previously is likely due to the difficulty of obtaining a sample size large enough to report a difference in diagnoses compared to a control population. Given that many of the studied conditions, like dyspareunia, endometriosis, leiomyoma, and uterine polyps, have a very low prevalence of less than 3% in our MFS cohort, research to assess the risk of these diagnoses needs a sufficiently sized population to observe smaller differences between groups [11, 12]. This is something we were able to capture due to our cohort size.

For urogynecologic conditions, we found an increased risk of urinary incontinence and pelvic organ prolapse in individuals with MFS. Therefore, this study adds to a small, but growing body of literature, that these conditions may actually be clinical manifestations of MFS [4, 5]. The sample sizes of previous studies included less than 20 individuals with MFS, compared to our cohort of 4347 MFS patients. Of considerable note, despite these urogynecologic conditions being typically associated with postmenopausal gynecologic care, our study found that there is an increased risk of urinary incontinence and pelvic organ prolapse even in MFS patients of reproductive age [13]. The role of TGF-β has been explored in these conditions as well, with dysregulation of TGF-β being documented in a murine model of stress urinary incontinence when increased vaginal distension is present [14]. Higher TGF-β expression has also been seen in severe cases of pelvic organ prolapse in postmenopausal women compared to moderate/mild cases [15].

There is a dearth of literature on the impact of MFS on female infertility. In our study, we documented an increased incidence of fertility testing in the MFS cohort compared to the control cohort. However, whether this testing was completed due to an actual concern for infertility, a desire to undergo ART for the opportunity to utilize preimplantation genetic testing for MFS in embryo selection, or a desire for fertility preservation is unable to be ascertained through TriNetX. Additionally, based on the potential impact of TGF-β dysregulation on the female reproductive system, we expected to see an increased risk of undergoing infertility treatment for MFS patients as well. While the absolute risk of infertility-related interventions was elevated in our MFS cohort, there were too few patients with each outcome to determine significance, likely due to underreporting of relevant ICD-10 codes. Thus, it is not out of the question that MFS patients may require more fertility intervention than patients without the condition, and potentially these patients may have been able to achieve pregnancy with a mechanism like intrauterine insemination, not needing to proceed to in vitro fertilization. We would expect a higher utilization than a control population at least in part due to a desire for preimplantation genetic testing for monogenic conditions (PGT-M) to determine whether an embryo has MFS [16]. Embryologists, physicians, and patients could consider the PGT-M results when selecting an embryo for transfer.

Despite the majority of MFS literature on women’s health being focused on obstetric outcomes, these studies largely focus on cardiac complications associated with pregnancy [3]. In our study, we found an increased risk for cesarean delivery, gestational hypertension, and pre-eclampsia. The increased risk for cesarean delivery has been previously reported (in a sample of 339 MFS patients), and we documented the same relative risk in our cohort (of 500 MFS patients) (1.91 vs 1.85), which provides further validation of our study approach [17]. The increased cesarean delivery in this population may be linked to adverse cardiac outcomes, like aortic root dilation and dissection, which patients with MFS are inherently at an increased risk for [18]. There is a lack of literature, however, on the incidence of gestational hypertension in individuals with MFS, but an increased risk for pre-eclampsia has been documented by Groth et al. but not corroborated by Baradwan et al. and Ramage et al. [19–21]. We found an elevated risk for both these diagnoses, which was unexpected because pregnant MFS patients typically are under strict blood pressure control to reduce the risk of aortic dissection [3]. It was interesting that both of the elevated blood pressure diagnoses we assessed were increased in the MFS cohort, indicating that something, like the need for increased blood pressure control or monitoring, may be contributing to these findings. Thus, more large-scale research in this area is needed to confirm the increased risk for these diagnoses in pregnant MFS patients. Finally, previous literature has documented an increased risk for PPH in MFS patients in a study of 29 pregnancies [22]. An increased risk of PPH may exist in this population, but we are unable to make that conclusion from our data due to a lack of statistical significance.

The rarity of certain genetic conditions often limits our ability to investigate associations with other diagnoses. Our study leveraged data from a large research network which enabled us to discover an increased risk for several gynecologic and obstetric diagnoses in a cohort of females of reproductive age with MFS. These findings can be used in conjunction with previous studies to counsel MFS patients on a potentially increased risk of the development of these conditions, with potentially profound implications for their future health and pregnancies. Further research is needed to prospectively assess the risk of these diagnoses in individuals with MFS.

This study provides compelling evidence to pursue translational studies on TGF-β dysregulation to determine the pathologic mechanism that leads to the development of these gynecologic conditions. Additionally, more research is needed to further elucidate the menstrual irregularities experienced by individuals with MFS. This includes the regularity/irregularity of menstrual cycles, the level of pain experienced during menstruation, and the quantity of menstrual bleeding. Prospective clinical studies of patients with MFS, such as the risk of undergoing gynecologic or urogynecologic surgeries, the response of MFS patients to fertility treatment, the use of PGT-M for embryo selection in this population, and whether the risks of hypertensive disorders of pregnancy can be mitigated in this population are also warranted. From a basic and translational perspective, patients with MFS provide a unique opportunity to study TGF-β signaling in the reproductive tract, with the opportunity for advanced transcriptomics and proteomics in lesions and tissues of MFS patients with endometriosis and fibroids.

Strengths and limitations

While the large sample size generated using a large database is a strength of this study, it is not without its limitations. One limitation of this study, largely conferred by the use of the TriNetX database, is the reliance on coding for analysis, meaning that providers may not have coded all diagnoses, properly coded diagnoses and that the codes themselves are not all-inclusive. It is possible that MFS patients are more likely to receive an increased number of diagnoses by care providers because of increased health care utilization due to their genetic condition. Of note, we tried to limit this error by including patients in both cohorts who have had an encounter within the health care system in the past ten years, meaning that the patients in the control cohort are more likely to represent a population that interacts with the health care system. The diagnosis of MFS itself may also be misreported through ICD-10 coding, as there is no genetic testing available to confirm these diagnoses. An additional limitation of this study is that it is a retrospective, observational study. Therefore, we cannot attribute causation to the reported findings.

Conclusions

Our study provides compelling evidence that individuals of reproductive age with MFS suffer from several gynecologic and obstetric conditions at greater frequency than their age-matched counterparts. Many of these diagnoses involving the female reproductive tract have not been part of the traditional counseling and care of MFS patients. This study further demonstrates a great need for additional basic, translational, and clinical research into MFS and the role of TGFβ signaling in the female reproductive tract.

Data availability

Data will be made available to the editors of the journal for review or query upon request.

Declarations

Conflict of interest

The authors declare no competing interests.