Current progress in clinical, molecular, and genetic aspects of adult fibromuscular dysplasia

Abstract

Fibromuscular dysplasia (FMD) is a non-atherosclerotic vascular disease that may involve medium-sized muscular arteries throughout the body. The majority of FMD patients are women. Although a variety of genetic, mechanical, and hormonal factors play a role in the pathogenesis of FMD, overall, its cause remains poorly understood. It is probable that the pathogenesis of FMD is linked to a combination of genetic and environmental factors. Extensive studies have correlated the arterial lesions of FMD to histopathological findings of arterial fibrosis, cellular hyperplasia, and distortion of the abnormal architecture of the arterial wall. More recently, the vascular phenotype of lesions associated with FMD has been expanded to include arterial aneurysms, dissections, and tortuosity. However, in the absence of a string-of-beads or focal stenosis, these lesions do not suffice to establish the diagnosis. While FMD most commonly involves renal and cerebrovascular arteries, involvement of most arteries throughout the body has been reported. Increasing evidence highlights that FMD is a systemic arterial disease and that subclinical alterations can be found in non-affected arterial segments. Recent significant progress in FMD-related research has led to improve our understanding of the disease’s clinical manifestations, natural history, epidemiology, and genetics. Ongoing work continues to focus on FMD genetics and proteomics, physiological effects of FMD on cardiovascular structure and function, and novel imaging modalities and blood-based biomarkers that can be used to identify subclinical FMD. It is also hoped that the next decade will bring the development of multi-centred and potentially international clinical trials to provide comparative effectiveness data to inform the optimal management of patients with FMD.

Graphical Abstract

Current understanding of FMD.

Current understanding of FMD.

1. Introduction and aim of the review

Fibromuscular dysplasia (FMD) is a non-atherosclerotic vascular disease that may involve medium-sized muscular arteries throughout the body (Table 1). The frequency of FMD in the general population is not known. Though FMD may occur in men, the majority of FMD patients are women (82–95% in contemporary registries) (Table 2).^1,2 FMD can occur at any age, including children and elderly patients, though presentation in young and middle adulthood is most common. Mean age at diagnosis of FMD is estimated as 43–53 years across different registries (Table 2). The pathogenesis of FMD remains largely unknown, but is probably multifactorial including genetic and environmental factors. However, it is not thought to be an inflammatory process (i.e. vasculitis) and it is also not known to involve the venous system. The angiographic features of FMD include the classical multifocal (string-of-beads) lesions as well as tubular or focal stenosis (now grouped under the generic term ‘focal’)^1,2 (Figure 1 and Table 1). These findings may be associated with the clinical manifestations of the disease depending on the stenosis-induced ischaemia of the vascular territory predominantly involved. While FMD most commonly involves the renal (66–91%), cerebrovascular (25–80%), and less often mesenteric (14–21%) and lower extremity (10–45%) arteries, involvement of most arteries throughout the body has been reported (Table 2).

Table 1.

Consensus points from the International FMD Consensus—Definitions and Nomenclature for FMD. Adapted from Refs¹,²

FMD is classified based on angiographic evaluation as multifocal or focal FMD (Figure 1).

To confirm FMD diagnosis ≥1 multifocal or focal arterial lesion must be present. Aneurysm, dissection, or tortuosity without multifocal/focal lesion/s is not sufficient to diagnose FMD.

In patients with multifocal/focal lesion/s in ≥1 vascular bed, if aneurysm, dissection, or tortuosity is present in another/other vascular beds a multivessel involvement of all affected vascular beds should be considered.

Table 2.

Comparison of selected data from the US, French (ARCADIA), European/International FMD registries and ARCADIA-POL study

Registry	US Registry for FMD (USA)1	European/International Registry3	ARCADIA (France, Belgium)4	ARCADIA-POL (Poland)5
Number of patients analysed, n	1885	1022	469	232
No. of centres evaluating patients	13	46	16	1a
Women (%)	95	82	84	83
BMI (kg/m2)	26	25	24	25
Family history of FMD (%)	5.4	3.0	2.4	2.6
Age at diagnosis of FMD (years)	53	46	53	43
Hypertension (%)	67	86	77	91
Office blood pressure (mmHg)	132/75	140/85	139/83	134/83
Age at hypertension diagnosis (years)	45	37	No data	36
Multifocal FMD (%)	95c	72	92	82
Multivessel FMD (%)	55.1	57.4	48.0	30.2
FMD lesions distribution
Renal (%)b	66	91	79	88
Cerebrovascular (including extracranial) (%)b	80	63	50	25
Mesenteric (%)b	15	21	17	14
Lower extremity (%)b	45	31	15	10
FMD-associated vascular complications
Dissections (%)	28	6	16	13
Aneurysms (%)	23	22	26	31

BMI, body mass index; FMD, fibromuscular dysplasia; IQR, interquartile range.

^aA total of 32 centres are actively recruiting patients and referring them to one national center.

^bRate in patients in whom evaluation was performed.

^cExcluding patients with undetermined/missing FMD phenotype, in addition 2% of patients had both multifocal and focal FMD.

Figure 1.

Multifocal (A) and focal (B) lesion of FMD of the renal arteries.

Extensive prior studies have correlated the arterial lesions of FMD to histopathological findings of arterial fibrosis, cellular hyperplasia, and distortion of the abnormal architecture of the arterial wall.⁶ More recently, the vascular phenotype of lesions associated with FMD has been expanded to include arterial aneurysms, dissections, and tortuosity. However, in the absence of a string-of-beads or focal stenosis, these lesions do not suffice to establish the diagnosis.^1,2,7 The scope of this review is to present an outline of recent advances and ongoing progress in the understanding of FMD and related diseases, made since publication of the International FMD Consensus^1,2 or not extensively covered in this document, from bench to bedside. A comprehensive summary of clinical management based on the International FMD Consensus^1,2 can be found in Supplementary Table S1.

2. Progress in research on FMD—multicentre registries

2.1 US registry for FMD

The US Registry for FMD, launched in 2009,⁸ currently includes nearly 3000 patients from 17 specialized FMD clinical centres (Figure 2 and Table 2). It is coordinated by the University of Michigan’s Michigan Cardiovascular Outcomes Research and Reporting Program and funded by the patient advocacy organization the FMD Society of America (www.fmdsa.org). New centres have been incrementally added across the USA to increase geographical, racial/ethnic, and clinical diversity of the Registry population. Key reports of the Registry published since 2012 include: description of the most common clinical manifestations of FMD, including hypertension, headache,⁹ and pulsatile tinnitus;⁸ the finding of cerebrovascular (carotid and vertebral) involvement of FMD being as common as renal involvement;⁸ a report of a high prevalence of aneurysms and dissections among patients with FMD;^10,11 a description of a higher prevalence of aneurysms and dissections in men with FMD,¹² and of a more benign vascular phenotype in patients diagnosed ≥65 years old.¹³

Figure 2.

National and international registries evaluating patients with FMD: The United States Registry for FMD (green), The French-Belgian ARCADIA network (orange/blue), Polish ARCADIA-POL Registry (red), and the European/International FMD Registry (blue).

2.2 ARCADIA registry

The French-Belgian ARCADIA Registry⁴ represents the ultimate development of a series of French achievements on FMD mostly produced by the Hypertension Excellence Centre of the Hôpital Européen Georges Pompidou (Paris, France).^14,15 It includes expert centres from 16 university hospitals in France and Belgium (Figure 2). The ARCADIA registry is at the origin of the landmark ARCADIA study⁴—which provided definitive evidence of a high prevalence of multivessel FMD based on systematic state-of-the art imaging of brain-to-pelvis main arterial beds (by CT- or MR-angiography) with centralized image reading in 469 patients (Table 2). It also provided the backbone for the PROFILE study, looking at the incidence of novel FMD localizations and complications during a 3 year-follow-up in patients who had been fully examined at baseline. The results of the PROFILE study are expected to be published in 2021.

2.3 ARCADIA-POL study

The Polish study ARCADIA-POL initiated in January 2015 at the National Institute of Cardiology in Warsaw currently includes more than 300 patients with confirmed FMD referred by 32 centres (Figure 2). All patients are examined by head to pelvis CT angiography and benefit from an exhaustive work-up during a 5 day hospitalization. As of now, the main achievement of the ARCADIA-POL study was to document a high prevalence of previously undetected FMD lesions affecting clinical decisions using this systematic approach in the first 232 patients (Table 2).⁵ Other contributions include evaluation of structural and functional changes in the heart and kidneys of patients with FMD,^16,17 and description of prevalence and characteristics of spontaneous cervical artery dissections¹⁸ and visceral FMD.¹⁹

2.4 European/International FMD registry

The European/International FMD registry was launched in Brussels in December 2015^3,20 and endorsed by the European Society of Hypertension (ESH) in 2016. It forms the nucleus of the wider European/International FMD Registry and Initiative (FEIRI).³ As of now, it enrolled about 1800 patients from 47 centres in 23 countries, also including Argentina, China, and Japan (Figure 2D). Currently, the main achievements of the European/International FMD Registry are (i) detailed analysis of the first 1000 patients included in the Registry, allowing characterization of distinct patient profiles according to FMD subtype, age and gender and identification of novel predictors of multivessel disease, aneurysms, and dissections³ (Table 2) and (ii) first-time detailed assessment of the prevalence and nature of FMD-related complications during pregnancy.²¹

3. Genetics of FMD

While both sporadic and familial forms of FMD exist, symptomatic FMD in relatives is reported only in a minority of cases (<5%). In some families, autosomal dominant inheritance with incomplete penetrance has been suggested.^22,23 Traditional, family-based genetic studies have been challenging to conduct due to the relatively low frequency of well-characterized multiplex pedigrees and incomplete penetrance (∼0.5).^23–26 FMD is currently thought to have at least a partially complex genetic basis, due to the influence of common genetic variants. Additionally, studies on genetic background of FMD are difficult due to the likely high prevalence of asymptomatic FMD patients (∼3–6%)^8,27 and the influence of potential environmental modifiers (e.g. female hormones, lifetime mechanical stress, and tobacco smoking).

Using a genome-wide association study approach, a single nucleotide polymorphism (SNP) rs9349379-A, in the Phosphatase and Actin Regulator 1 (PHACTR1) gene has been identified as a common genetic risk variant. The protein encoded by this gene is a member of the phosphatase and actin regulator family of proteins. This family member can bind actin and regulate the reorganization of the actin cytoskeleton. It plays a role in tubule formation and endothelial cell survival.²⁸ The associated SNP confers an odds ratio of ∼1.4 for FMD.²⁹ It is located within an intron of the PHACTR1 gene and is associated with PHACTR1 transcript expression levels in dermal fibroblasts.²⁹ The same variant has been associated with spontaneous coronary artery dissection (SCAD) and expression QTL and colocalization analyses in this context have supported an effect of this SNP on PHACTR1 expression in vascular tissues.^30,31 Additional data suggest that this SNP is located at the site of an enhancer in vascular tissue, where it has been shown to regulate PHACTR1 and endothelin-1 (EDN-1) expression.³² The potential involvement of EDN-1 is especially intriguing as it has several known vascular effects on arterial tone and remodelling. However, at this time, the precise pathological changes in vascular function resulting from the PHACTR1 risk variant have yet to be defined.³³ Notably, this risk locus has pleiotropic associations, with SCAD, carotid artery dissection, hypertension, and migraine headache, as well as an inverse association with coronary atherosclerotic arterial disease and related myocardial infarction^{30,31,34–39} (Figure 3). Among individuals with FMD, a recent analysis of a genetic risk score for SCAD identified increased risk of SCAD among individuals with FMD, supporting that while these phenotypes overlap clinically and genetically at loci such as rs9349379, there are distinct genetic risk factors as well.³⁰ It is noteworthy that FMD, SCAD, and related diseases all have a higher prevalence in women, typically manifesting at younger ages than the timeframe of disease manifestations from atherosclerotic arterial disease,⁴³ which suggests possible shared genetic and non-genetic risks.

Figure 3.

Association of PHACTR1 variant with FMD, spontaneous cervical or coronary artery dissection, coronary artery diseases, and related diseases. (i) Warchol-Celinska et al.—updated analysis of previously published data by Kiando et al.⁴⁰ based on 1283 FMD cases and 4193 controls;⁴¹ (ii) Debette et al.—based on 1393 SCeAD cases and 14 446 controls;³⁸ (iii) Adlam et al.³¹—based on 1055 SCAD cases and 7190 controls; (iv) Nikpay et al.³⁶ based on 43 171 cases and 127 176 controls; (v) Gupta et al.³² based on Nikpay et al.³⁶ and UK Biobank, 65 262 cases and 231 381 controls; (vi) O’Donnell et al.;⁴² and (vii) Gupta et al.³² based on Anttila et al.³⁹ and UK Biobank, 7995 cases and 116 409 controls. OR, odds ratio; FMD, fibromuscular dysplasia; SCAD, spontaneous coronary artery dissection; SCeAD, spontaneous cervical artery dissection.

Notably, in the European/International FMD Registry, the overall prevalence of atherosclerotic lesions reported by the investigators was 17% (171/1012).³ While, expectedly, it was almost double in a subset of 122 patients aged 65 or older (33%), this prevalence remains surprisingly low for age. As a matter of comparison, in the Cardiovascular Health Study, including 5888 participants aged ≥65 years, the prevalence of carotid atherosclerotic plaques was 77%.⁴⁴ In agreement with the lower prevalence of cardiovascular events in elderly patients from the US Registry,¹³ this may reflect survival or inclusion biases. Alternatively, it may support the belief that patients with FMD are somehow protected from atherosclerosis.

A number of rare, exonic coding genetic variants have been implicated in FMD. Exome sequencing has identified loss-of-function variants in PTGIR, which encodes for the prostaglandin I2 receptor.⁴⁵ Recently, a recurrent novel variant in the collagen type V alpha 1 chain COL5A1 gene, c.1540G>A, p.(Gly514Ser), resulting in the substitution of a glycine residue with a serine residue at position 514 of the protein, has been associated with a phenotype encompassing multifocal FMD, arterial dissections, aneurysms, and tortuosity.⁴⁶ This finding is the first clinically actionable genetic finding for adult multifocal FMD. The phenotype is notable for involvement of the external iliac artery and coeliac artery with arterial aneurysms and dissections, and carotid artery tortuosity was observed as well. Spontaneous arterial dissections were seen in individuals harbouring this variant in the coronary, carotid, coeliac, and iliac arteries. Clinical presentations occurred in the fourth or fifth decades of life, and proposed management was similar to that for vascular Ehlers–Danlos syndrome due to collagen type III alpha 1 chain (COL3A1) pathogenic variants, including consideration of surveillance angiographic imaging, blood pressure monitoring and control, consideration of pregnancy risk, avoidance of contact sports, and surgical precautions. The recurrent COL5A1 variant is rare, as it was not observed in genetic population databases, such as gnomAD. Further, the variant was found to exist on a shared ancestral haplotype, supporting a ‘founder effect’ that would implicate additional individuals in the population harbouring the variant. Finally, additional individuals with multifocal FMD also harbouring genetic variants in the COL5A1 gene that were predicted to be deleterious by in silico analysis were predisposed to arterial dissections.⁴⁶ Whether the additional variants in this gene are pathogenic or may be acting as modifiers of FMD remains to be clarified.

Future directions of FMD genetic investigation include comprehensive genome-wide analyses of genetic variation including both rare, high-impact alleles hypothesized to underlie familial forms of FMD, as well as common genetic variation contributing to a complex genetic architecture of FMD. Cutting-edge genetic technologies have recently made these types of investigations feasible, and accordingly, these studies are now underway to provide urgently needed genetic information for FMD.

4. Molecular studies in FMD

Over the last decades, increasing evidence has emerged that altered Transforming growth factor-β (TGF-β) signalling plays a crucial role in the pathogenesis of connective tissue disorders. A first observation was made in a mouse model of Marfan syndrome (MFS), a multisystemic disorder characterized by aortic aneurysm/dissection, lens dislocation, and skeletal overgrowth. TGF-β neutralizing antibodies administered to a MFS mouse model demonstrated that increased TGF-β signalling was driving the multisystemic aspects of the disease in addition to the historically known structural deficiency of fibrillin-1, an extracellular matrix protein that is mutated in MFS.^47,48 The important role of TGF-β signalling pathway was further supported by the identification of mutations in the genes encoding for components of this pathway, affecting the cytokines (TGFB2/3), the receptors (TGFBR1/2), and the downstream effectors (SMAD2/3) in a new aortic aneurysmal syndrome, now called Loeys–Dietz syndrome (LDS).^49–51

Clinical studies in patients with FMD indicated that 30–57% of patients have angiographical manifestations in multiple arterial beds.⁸ In addition, some FMD patients may also have connective tissue features beyond the arterial pathology, including low bone density, joint laxity, scoliosis, early onset arthritis, and degenerative spine disease.⁵² Based on these observations, FMD is increasingly considered as a systemic disease with some overlap with known connective tissue disorders, including MFS, LDS, and Ehlers–Danlos syndrome. As altered TGF-β signalling was already involved in these conditions, a role for TGF-β in the pathogenesis of FMD was hypothesized.⁵² In a small cohort of 38 FMD patients, it was shown that both circulating TGF-β1 and TGF-β2 plasma levels were increased.⁵² In fibroblasts of FMD patients, no effect on downstream TGF-β pathway effectors could be demonstrated, even when stimulated with TGF-β1, TGF-β2, or bone morphogenetic protein 4.⁵² Overall, these findings suggest that activation of the TGF-β pathway is not a primary cause of FMD. This is also confirmed by the fact that mutational screening of TGF-β pathway related genes in FMD patients has yielded very few mutations. In the cohort of Ganesh et al.,⁵² no mutations were found in TGFBR1/2, TGFB2, and SMAD3 in 47 patients. In another series of 35 patients with FMD who underwent genetic evaluation, 2 variants (p.Thr204Ile & p. Tyr429His) were identified in TGFBR1.⁵³ Both patients had multiple cervical artery dissections and thoracic aortic aneurysm as well as more classical multifocal FMD lesions, so may represent LDS variants. No familial segregation was done for these two variants. A recent study⁵⁴ analysed all known pathogenic TGF-β signalling pathway genes in a cohort of 179 SCAD patients (with or without FMD) and 102 patients with FMD only. Whereas a mutational burden was identified in the SCAD cohort in the known LDS genes, the latter was not found in the FMD only cohort. Notably, in most cases, the distinction between FMD and MFS, LDS and Ehlers–Danlos syndrome is easy to make based on the presence or absence of typical string-of-beads arterial lesions. Though common pathogenic mechanisms are postulated, these are clearly separate entities.

Recently, researchers uncovered a possible link between FMD and mutations in a second pathway, driven by platelet-derived growth factor (PDGF) receptor β (encoded by the PDGFRB gene). PDGFRB is highly expressed in vascular mural cells and drives their migration and proliferation in angiogenesis.⁵⁵  PDGFRB gain-of-function mutations cause myofibroma, classified as a benign pericytic tumour made of myofibroblastic cells.^56,57 The development of multiple myofibroma defines infantile myofibromatosis, a disease that may become life threatening when internal organs are involved. Most of the mutant receptors identified in myofibromatosis are sensitive to tyrosine kinase inhibitors, such as imatinib in vitro, which was successfully tested in several severely affected patients.^56,58

In 2010, Brasseur et al.⁵⁹ reported myofibromatosis also involving renal arteries in a young patient with severe hypertension, multiple myofibromas and aneurysms of the renal and iliac arteries. Sequencing tumour DNA revealed an acquired PDGFRB p. D850V substitution,⁵⁷ which was also present in the renal artery lesion (Dachy and Demoulin, personal communication). This mutation constitutively activates the receptor signalling in the absence of ligand.⁵⁷ Recently, somatic activating mutations were also identified in four out of six cases of cerebral fusiform aneurysms.⁶⁰ Interestingly, several of these PDGFRB variants were identical to those found in myofibroma. The association between PDGFRB and aneurysms was further confirmed in patients with Kosaki overgrowth syndrome, another rare connective tissue disease caused by germline PDGFRB variants. Two patients were reported with fusiform aneurysms of the basilar artery, resulting in a lethal haemorrhage at the age of 20, and stroke, respectively.⁶¹ In another report, saccular aneurysms of coronary arteries and arterial tortuosity were diagnosed in a patient with Kosaki syndrome who suddenly died at the age of 19.⁶² Altogether, these studies confirm that overactivation of PDGF receptor β causes aneurysms and, possibly, other vessels anomalies. Nevertheless, whether PDGFB and PDGFRB genes are involved in the pathophysiology of classic FMD remains to be demonstrated.

5. Proteomics and transcriptomics of FMD

It has been known for several decades that FMD involves changes in the composition of the arterial wall.^63,64 However, unlike diseases such as atherosclerosis where samples from affected vessels can be obtained via surgery, the vast majority of patients with FMD are treated conservatively or with catheter-based therapy (e.g. angioplasty). As a result, FMD vascular tissue samples are very rarely obtained. As a more novel line of investigation, there have been increasing efforts directed towards leveraging blood and dermal fibroblast samples to gain insights on the pathophysiology of FMD. Notably, dermal fibroblasts can be readily obtained from a small (2–3 mm) skin biopsy using outgrowth tissue culture techniques over 4–8 weeks, and it is intuitive that FMD involves alterations in fibroblast cells. Indeed, as mentioned, Ganesh et al.⁵² found elevated plasma TGF-β and inflammatory marker levels, and increased TGF-β secretion in dermal fibroblast cell lines from subjects with FMD compared to age- and gender-matched controls.⁵²

More recently, Olin et al.⁶⁵ published the first report from the DEFINE-FMD study, a functional multi-omics systems biology study aiming to decipher the molecular and genetic basis of FMD. Of relevance, systems biology approaches leverage highly tractable and informative methods⁶⁶ that have been used with success for understanding coronary artery and other vascular diseases.^67,68 In their study, Olin et al.⁶⁵ evaluated plasma levels of 981 proteins and 31 lipid sub-classes in women with multifocal FMD and matched healthy controls. Using separate discovery and validation cohorts with a combined total of 113 FMD cases and 128 controls, they identified and successfully validated a signature of 37 plasma proteins and 10 lipid sub-classes with differential abundance in FMD patients vs. controls. Using systems biology approaches, they showed that the genetic locus of one of these ‘FMD signature proteins’, CD2-associated protein (CD2AP), was independently associated with risk of having FMD (P=0.0003) and that it is widely expressed by endothelial cells in medium-large-sized human arteries (Figure 4). In addition, machine learning trained on the discovery cohort plasma protein and lipid data were used to develop a test for FMD. When independently applied to the validation cohort, the test showed promising proof-of-principle for its ability to diagnose the presence of FMD. Accordingly, an accompanying editorial suggested that while these exciting findings require further validation, the potential clinical relevance of a blood-based test for FMD may include the ability to: (i) predict the presence of silent FMD in relatives of patients with symptomatic FMD; (ii) support or rule out the diagnosis of FMD in patients with suggestive arterial features but without the pathognomonic ‘string-of-beads’; and (iii) predict progression of the disease.⁶⁹

Figure 4.

FMD signature proteins. CD2AP was associated with FMD and was expressed by endothelial cells. Immune-fluorescence staining for CD2AP is performed on adult human non-FMD samples from renal artery (A), internal mammary artery (B), and aorta (C). Endothelial cells are identified by staining for CD31 (green), while CD2AP is shown in red. Nuclei are stained with DAPI (blue). Scale bar—25 µm. Inset panels on the right, represent a magnified view of the area in the respective dashed squares, show endothelial cells at higher enlargement. M—tunica media; L—lumen. Reproduced with permission from Olin et al.⁶⁵

DEFINE-FMD has now enrolled over 400 subjects and while unpublished, additional studies using the fibroblast samples from this study appear equally promising. Again using systems biology approaches but on this occasion applied to transcriptomic (i.e. gene expression) data from fibroblasts, the investigators have reportedly identified a gene regulatory network that is of importance in FMD, which has led to their developing mouse model of this disease.⁷⁰

6. Immunological and inflammatory aspects of FMD

The current definition of FMD emphasizes its fibrotic, non-atherosclerotic, and non-inflammatory nature. This last statement is based on the early observations of relative paucity of inflammatory cells in advanced FMD lesions at a time where vascular samples were available. However, in view of the critical role of inflammation in triggering vascular fibrosis, which has emerged in recent years, including the role of key cytokines, such as TGF-β, IL-17, and IL-9 in the regulation of fibrosis,^52,71 it may be hypothesized that inflammation is involved at the early stages of FMD. In a recent study, angiotensin II was reported to lead to severe perivascular and vascular fibrosis in the absence of T cells and T cell-dependent cytokines. Moreover, FMD patients showed features of increased inflammatory biomarkers including tumour necrosis factor alpha, C-reactive protein, monocyte chemoattractant protein 1, serum amyloid A protein, intercellular adhesion molecule 1, vascular cell adhesion molecule 1 and IL-8, as well as clearly profibrotic TGF-β compared to control subjects and patients with MFS.⁵² Finally, the proteomic signature of FMD confirmed that number of top newly identified FMD-specific proteins are inflammatory.⁶⁵ These include CCL11 (eotaxin), fibroblast growth factor 19, or GRB2 related adaptor protein 2. Interestingly similar sets of mechanisms are linked with immune and inflammatory pathogenesis of hypertension itself.⁷² Recent studies using optical coherence tomography demonstrate that vascular segments affected by coronary FMD display features usually associated with vascular inflammation, such as multiple areas of patchy or diffuse intimal, medial or adventitial abnormalities, and features of macrophage infiltration.⁷³ Further mechanistic studies will be critical to provide detailed insights into the role of inflammation in the pathogenesis of FMD.

7. New and revisited risk factors for FMD

As discussed above, although a variety of genetic, mechanical, and hormonal factors may play a role in the pathogenesis of FMD, overall, its cause remains poorly understood. It is probable that the pathogenesis of FMD is linked to a combination of genetic and environmental factors e.g. smoking, mechanical factors, stretching, and maybe vasculotoxic medications, such as fluoroquinolones.

Tobacco smoking may be a potential pathogenic factor associated with FMD. Following early observation of Nicholson et al.⁷⁴, Savard et al.¹⁴ showed that the proportions of current smokers (30% vs. 18%) or ever smokers (50% vs. 37%) were much higher in FMD patients as compared to matched hypertensive controls (Figure 5). Moreover, the prevalence of current smoking was much higher in patients with focal than with multifocal lesions (50% vs. 26%, respectively), possibly reflecting the difference in sex distribution (female: male ratio, 2:1 and 5:1, respectively). Current smoking was also associated with a more aggressive course among patients with multifocal FMD. In the US Registry, history of smoking was associated with a significantly higher rate of aneurysm detection in patients with FMD (25% vs. 19% in never smokers) (Figure 5) as well as an increased prevalence of major vascular events (43% vs. 37%, respectively).⁷⁵

Figure 5.

Frequency of smoking in FMD patients in Savard et al., Dobrowolski et al., and O’Connor et al. studies.^14,75,76 HT, hypertensive; FMD, fibromuscular dysplasia.

Contrasting with these two studies, Dobrowolski et al. documented no difference in the rate of ever smokers among FMD patients as compared to two matched control groups—the first from the general population and the second consisting of hypertensive subjects (Figure 5). Moreover, smoking was neither associated with clinical characteristics of patients with FMD nor with extent or complications of the disease.⁷⁶ In summary, data linking smoking with the pathogenesis of FMD are equivocal and further studies are needed to elucidate this potential relationship.

The use of fluoroquinolones is associated with thoracic and abdominal aortic aneurysms or dissections.^77,78 Also, the use of fluoroquinolones was associated with a more than two-fold increased risk of spontaneous cervical artery dissection.⁷⁹ However, a recent study did not provide evidence for an excess of risk of intracranial aneurysm or dissection with fluoroquinolone use.⁸⁰ Experimental studies have demonstrated several molecular mechanisms responsible for fluoroquinolone-associated collagen toxicity, including increased matrix metalloproteinases (MMPs) activity and collagen degradation, as well as decreased activity of tissue inhibitors of MMPs.^81,82 Although there are no data on the role of fluoroquinolones in the pathogenesis of FMD and associated dissection, while waiting for specific evidence, patients with FMD should follow the advice of the FDA that ‘fluoroquinolones should not be used in patients at increased risk [of aortic ruptures or dissections] unless there are no other treatment options available’.⁸³

8. Arterial structure and function in FMD

Increasing evidence highlights that FMD is a systemic disease, and that subclinical alterations can be found in non-affected arterial segments, such as the ‘triple signal’ pattern in the common carotid artery^22,84 (Figure 6), or the presence of reduced arterial diameter and impaired smooth muscle cell function in the brachial artery.⁸⁵ Triple signal was associated with increased stiffness of the vascular wall of the common carotid artery⁸⁵ and may represent the ultrasound signature of outer media fibrosis observed in renal biopsies.²² However, carotid triple signal has limitations: it is less FMD-specific than previously thought, being also associated with atherosclerosis and traditional risk factors,^85,86 and it has not been validated in independent cohorts yet. Indeed, subclinical abnormalities seem to be more evident in the muscular medium-sized arteries (usually spared by atherosclerosis but specifically affected by FMD), according to a ‘muscular-to-elastic’ gradient.⁸⁵

Figure 6.

Subclinical ultrasound FMD features in non-affected districts include disarray of the arterial wall. Normal common carotid artery wall (standard ultrasound) (A) and normal radial artery (UHF-US) (B) are characterized by one and two echogenic interfaces, respectively (white arrows). In FMD patients, common findings are the presence of additional echogenic interfaces (yellow arrows), either in the posterior wall of the carotid artery (the so-called triple signal) (C) or in the radial artery (D).

The recently introduced technique of ultra-high-frequency ultrasound (UHF-US) allows imaging of superficial tissues (2 cm depth max), making it possible to study arterial wall thickness and ultrastructure of medium-sized and small-sized arteries (lumen down to 300 μm) in vascular diseases and therefore has the potential to overcome these limitations.^87–89

In 2017, the first research programme for non-invasive vascular phenotyping by UHF-US in diseases from the FMD-spectrum, the FUCHSIA study—‘Very high-Frequency Ultrasonography for arterial phenotyping in patients with Cervico-Cerebral Artery Dissection (CCeAD), Hypertension, Spontaneous Coronary Artery Dissection (SCAD) and FibroMuscular Dysplasia (FMD)’—was launched in four different countries (Italy, UK, France, and Belgium). Preliminary results from the Italian cohort showed an altered ultrastructure in carotid and radial walls of patients with FMD, with thickening and disarray of wall layers in comparison to healthy controls⁹⁰ (Figure 6).

In the next years, the FUCHSIA research plan includes replication in independent validation cohorts; integration of the radiomics—machine learning approach for classification; and correlation of vascular phenotypes with disease characteristics and clinical outcomes. Coupling with other omics techniques is also advisable,⁶⁹ for a more accurate diagnosis and risk stratification, as well as for a better understanding of the pathophysiology of FMD.

9. Renal haemodynamics in FMD

Until recently, little was known on the impact of multifocal renal artery FMD on the kidney. Most knowledge was based on experiments in animal models with clipping of the renal artery and studies in patients with atherosclerotic renal artery stenosis (ARAS). Recent studies in patients with renal artery FMD, however, shed a new light on the kidney in FMD and suggested important differences with ARAS and animal models.

First of all, it has been proposed that in multifocal FMD the intrarenal vasculature is relatively intact. Renal blood flow (RBF) was indeed measured in 64 patients with multifocal FMD (prior to balloon angioplasty and without the use of antihypertensive drugs) using the ¹³³Xenon washout method. RBF in the affected kidney was comparable to that in the unaffected contralateral kidney.⁹¹ Moreover, the global RBF in multifocal FMD was comparable to that in matched patients with essential hypertension, and significantly higher than that in patients with ARAS.^91,92

The ARCADIA-POL study evaluated intrarenal haemodynamic patterns by Doppler-ultrasonography in 153 patients with renal artery FMD.¹⁷ Among FMD patients with non-significant renal FMD, the renal resistive index (RRI) was comparable to that in patients with essential hypertension and healthy normotensives. Also, no difference in RRI was found between patients with multifocal and focal FMD. However, in FMD patients with significant stenosis, post-stenotic intrarenal RRI was significantly lower as compared to FMD patients without significant stenosis, hypertensive controls, and normotensive subjects.¹⁷ Likewise, lower RRI was measured in kidneys with FMD with a significant stenosis as compared to those without, along with significant differences in other intra-renal duplex ultrasound parameters.

Since the post-stenotic flow pattern is associated with the compliance of the post-stenotic vessel wall, a lower RRI may reflect a high compliance. This may explain the finding of a lower RRI in patients with FMD with significant stenosis.

Finally, microvascular function appears to be preserved in kidneys with multifocal FMD. Global renal function is generally in the normal range (mean eGFR in the European FMD registry was 92 mL/min/1.73 m²)³ and does not differ between the affected and unaffected kidney in patients with unilateral FMD if the renal artery stenosis is not severe.⁹¹ However, this may depend on the severity of stenosis.⁹³ Moreover, in patients with multifocal FMD enrolled in the ARCADIA-POL study, RRI was not correlated with renal function or albuminuria.¹⁷ Finally, intrarenal infusion of vasoactive substances in kidneys with FMD results in a haemodynamic response that is comparable to that in controls,⁹¹ while this response is severely decreased in ARAS (Figure 7).^94,95

Figure 7.

Kidney in FMD: tentative differences between FMD and ARAS. Adapted from Ref.⁹⁵

As a whole, it appears that the presence of a string-of-beads does not significantly impair renal perfusion or intrarenal microvascular function in most patients, which is in sharp contrast to ARAS. In the short term, this remains true even in the presence of a unilateral tight stenosis because of compensatory mechanisms in the contralateral kidney.⁹³ Presumably, the lower exposition to other pro-atherosclerotic and nephrotoxic factors (diabetes and hypercholesterolemia) and the shorter duration of hypertension, may also contribute to preserve the kidney tissue from glomerular and tubular damage and fibrosis, as well as intrarenal vascular compliance. Possibly, altered haemodynamics caused by the string-of-beads may also protect the kidneys from hypertensive damage. These differences in renal microvasculature may explain why revascularization is usually unsuccessful in ARAS^96,97 while it frequently has a blood pressure-lowering effect in FMD.⁹⁸

Renal perfusion is relatively preserved in patients with multifocal FMD and thus renin secretion is not markedly increased compared to patients with essential hypertension. Moreover, there is no renin secretion gradient between the affected and unaffected kidney in patients with unilateral multifocal FMD.^91,92 However, interpretation of these studies^4,17,95,99 is complex since in the absence of trans-stenotic pressure gradient measurements, a proportion of patients may have non-significant renal artery stenosis associated with essential hypertension.

In contrast, in patients with focal FMD, the impact on the kidney appears similar to that of ARAS and renal artery clipping in animal models: renal perfusion and glomerular filtration are lower in kidneys with focal FMD and renin secretion is increased in the affected kidneys.⁹⁹ Hence, it appears that the haemodynamic impact of focal FMD lesions on the kidney is more severe, resulting in a more classic pattern of renovascular hypertension. This may in turn explain the more often severe clinical presentation and higher success rate of balloon angioplasty in patients with focal FMD compared to multifocal FMD. Work regarding renal haemodynamics and haemodynamic assessment of renal artery FMD is ongoing. In 2019, the International FMD Consensus proposed a standardized clinical protocol for haemodynamic assessment of renal arteries based on trans-stenotic pessure gradient measurements prior to and following angioplasty, which requires additional validation.^1,2

10. Neurological manifestations of FMD

The most frequent presenting symptoms of neurological FMD are headaches and pulsatile tinnitus.^4,8,100 Headaches (mostly migraine type and tension type) were reported by 70% of patients with FMD.^4,8,9 The underlying pathophysiology of headaches in FMD patients remains vague and potential mechanisms comprise changes in cerebrovascular flow—labile hypertension, hyper- or hypoperfusion, neurovascular dysregulation or dysautonomia, structural injury—dissection, microtrauma, or enhanced pain sensitivity.¹⁰¹ Although headaches are commonly reported by FMD patients, irrespective of involved vascular beds, the presence of headaches is significantly more frequent in FMD patients with a history of cervical or intracranial artery dissection or intracranial aneurysm.⁹ Pulsatile tinnitus is reported in 40% of patients with neurological FMD and may be associated with the presence of cervical artery dissection.^8,9

Patients with FMD have a high prevalence of cervical (carotid and vertebral) artery dissection and intracranial saccular aneurysms.^10,102 The most frequent complications of neurological FMD are, in decreasing order, transient ischaemic attack (TIA) (8–53%) or ischaemic stroke (8–35%), subarachnoid haemorrhage or unruptured aneurysm (3–49%), and intracerebral haemorrhage (6–13%).¹⁰²

In FMD patients, TIA or ischaemic stroke can be a consequence of: haemodynamic mechanism (hypoperfusion from cervical or intracranial arterial stenosis or occlusion—Supplementary Figure S2A), emboli (thrombosis in the area of a stenosis or dilation), or thrombosis (lacunar stroke, thrombosis of small, perforating arteries due to secondary hypertension). However, TIA/ischaemic stroke appear to occur primarily in the presence of associated cervical artery dissection, due to artery-to-artery thromboembolism or cerebral hypoperfusion (Supplemental Figure S2B). Subarachnoid haemorrhage mainly results from the rupture of an intracranial aneurysm and less commonly from dissecting vertebral artery into the intracranial segment.¹⁰² Intracerebral haemorrhage is rare in patients with FMD but may be due to the rupture of an intracranial aneurysm, a dissection or the presence of hypertensive microangiopathy.¹⁰²

FMD is the most common underlying vasculopathy at the origin of cervical artery dissection, and a recent systematic review of patients with cervical artery dissection estimated the presence of FMD lesions to be 16%.¹⁰³ Also, in a recent small study of 43 patients with spontaneous cervical artery dissection who underwent brain-to-pelvis CT-angio evaluations, FMD was found in 39.5% of patients.¹⁸

The prevalence of cervical artery dissection has been estimated to be 21% in all FMD patients, with an increased prevalence of carotid (16%) as opposed to vertebral (5%) artery dissection.¹⁰⁴ Furthermore, among patients with a neurological presentation leading to the diagnosis of FMD, the prevalence of cervical artery dissection is even higher, being 27%.⁴ Cervical arteries remain the main location of arterial dissections in FMD patients, with cervical artery dissection accounting for up to 65% of all dissections.¹⁰⁴ Of note, among FMD patients with a cervical artery dissection, the presence of multiple cervical artery dissections is not uncommon and estimated to be up to 37%.^8,10

The ARCADIA registry has identified male sex, age >50years, history of migraine or hypertension and involvement of ≥3 vascular beds with FMD lesions to be independent risk factors for cervical artery dissection in patients with FMD (Arnaud et al., Journal of American Heart Association, in press).

Intracranial saccular aneurysms are mainly unruptured and diagnosed incidentally on imaging rather than discovered after rupture. The prevalence of intracranial aneurysms in patients with FMD is higher (7%) than in the general population (<5%).¹⁰⁵ In the US Registry, 13% of women with FMD had at least one intracranial aneurysm and 4% more than one. Of interest, 29% of these aneurysms were of size ≥5 mm, which is the threshold commonly judged to classify intracranial aneurysms as of high risk of rupture.¹⁰⁶ Also, there was no difference in the rate of intracranial aneurysms between patients presenting with renal and cervical FMD.¹¹ It remains unclear whether the risk of rupture of intracranial aneurysm is higher in FMD patients that in the general population (<1%/year), because of the shortage of longitudinal data in patients with FMD with intracranial aneurysm.

Information on risk of long-term progression of FMD and occurrence of non-specific lesions (aneurysm, dissection etc.) is scarce. Longitudinal data from the US FMD registry suggest that FMD is a non-progressive disease, with no or rare extension of FMD lesions in initially involved arteries and no FMD involvement in previously unaffected arteries.¹⁰⁴

Similarly, patients with cervico-encephalic FMD do not appear to have an increased risk of FMD progression and occurrence of non-specific lesions (e.g. recurrent cervical artery dissection).¹⁰⁴ Indeed, in 146 patients with multifocal FMD with follow-up cervical imaging, none had development of FMD in a previously unaffected artery and none had FMD progression in previously involved artery after a mean follow-up of 35 months (range 5–153 months).¹⁰⁴ No patient had a new aneurysm and three (2%) had a new cervical artery dissection during follow-up. Interestingly, all three patients with a new cervical artery dissection during follow-up already had multifocal FMD lesions on the same cervical artery at baseline, one had a previous contralateral cervical artery dissection and another a previous coronary artery dissection.

Arterial diaphragms are thin translucent endoluminal webs localized in the carotid (especially in the posterolateral side of the carotid bulb) or vertebral (V3 segment or ostium) artery and correspond to a linear defect on angiography that does not change or disappear after change of the patient’s head position (Supplemental Figure S2C).^107–109 This entity has been reported to share similar histological findings with FMD, with anatomical specimens showing intimal fibroplasia without atheromatous or inflammatory lesions in almost all cases.^102,110,111 Several authors estimate that this could be the predominant form of FMD in Black populations and have classified it as ‘atypical FMD’.¹¹²

Carotid webs have been reported in middle-aged adults, predominantly women, of African or Afro-Caribbean ethnicity, presenting with a TIA or ischaemic stroke in the carotid artery territory ipsilateral to the web, with no other causes identified after a comprehensive work-up. Patients with carotid webs have no typical multifocal or focal FMD lesions, no aneurysm or dissection, and no atherosclerotic risk factors.^{107–109,112} Carotid webs could be more frequent than previously thought in patients who suffered an ischaemic stroke. A review of consecutive patients with ischaemic stroke indeed reported a 1.2% rate of carotid web (0.7% ipsilateral to the stroke),¹⁰⁸ while a subanalysis of patients with ischaemic stroke due to intracranial large-vessel occlusion included in the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischaemic Stroke in the Netherlands (MR CLEAN trial) showed a 2.7% frequency (2.5% ipsilateral to the stroke).¹¹³

Cerebral ischaemia is thought to be mediated by an embolic mechanism because of stasis upstream to the web or within an associated aneurysmal bulb or from focal dissection, and a strong association between carotid webs and ischaemic stroke has been reported in a population-based case–control study.¹¹²

11. Heart and FMD

11.1 Heart morphology and function in FMD

Contrary to atherosclerotic RAS, data on the relationship between renal FMD and left ventricular morphology and function are limited.^114–116

In the ARCADIA-POL study, LV morphology and function of 144 FMD patients were compared to those in a matched control group. No significant differences in left ventricular morphology and systolic function were found between FMD patients and controls or between subjects with multifocal and focal FMD, even in case of significant RAS.¹⁶

In contrast, early alterations in left ventricular diastolic function (lower e’ and higher E/e’) were found in patients with multifocal FMD compared to both patients with focal FMD and controls.¹⁶ These results may be explained by the older age of subjects with multifocal FMD.

In summary, there is no evidence that FMD is associated with alteration in LV morphology and function over and above BP levels.

11.2 Coronary arteries in FMD

The coronary phenotype in patients with FMD may include the presence of excessive tortuosity, the intravessel or multivessel symmetry sign, corkscrew appearance, dissection, aneurysm, distal tapering or long, smooth narrowing in relation with either a dissection or an intramural haematoma within the coronary artery.^1,2,117

The most common feature of coronary phenotype in FMD is coronary tortuosity, however, the diagnostic criteria applied across studies are non-uniform and arbitrary (Supplementary Figure S1).^118–120

Coronary tortuosity has been reported in all of 32 examined patients with extra-coronary FMD in one study.⁷³ Similarly in another study five patients with resistant hypertension, exertional chest pain, and significant coronary tortuosity were diagnosed with FMD of renal and/or cervical arteries.¹²¹ Corkscrew appearance and multivessel symmetrical tortuosity were also associated with extra-coronary FMD in SCAD patients.¹²⁰ On the other hand, coronary tortuosity defined as ≥3 bends above 45° may be found in 39.1% of consecutive patients undergoing coronary angiography due to chest pain whereas it may be found in 12.5% when tortuosity is defined as ≥2 bends above 180°.^118,119 These results should therefore be interpreted with caution due to limited number of patients and potential selection bias.

Clinical presentation of coronary artery alterations observed in FMD ranges across a wide spectrum from asymptomatic, through exertional chest pain, to rarely life threatening, such as myocardial infarction or sudden cardiac death mostly due to SCAD. More severe coronary tortuosity is linked to an increased risk of recurrent coronary dissection in SCAD.¹²⁰ Coronary tortuosity may increase transcoronary gradient leading to myocardial ischaemia and increased local shear stress, which may facilitate dissections.^122–124

11.3 SCAD and FMD

SCAD is a cause of acute coronary syndromes, which, like FMD, afflicts predominantly young to middle-aged women. Following the first report of extra-coronary arteriopathies in SCAD-survivors,¹²⁵ a range of vascular abnormalities have been described including aneurysms, dissections, focal stenoses, and aortic root dilation but the commonest finding is an appearance indistinguishable from the typical radiological manifestations of multifocal FMD.^126,127 To date, there has been no histological confirmation that these appearances arise from the same changes at a cellular level as those described in classical FMD¹²⁸ but it is now widely accepted that the radiological findings in SCAD represent true multifocal FMD rather than a phenocopy.^129,130

It remains uncertain if all other vascular abnormalities seen in SCAD are part of the FMD-spectrum or if FMD is just one of a range of SCAD-associated arteriopathies. Whilst coronary arterial wall abnormalities have been described from intravascular imaging in SCAD,^73,131 typical multifocal FMD-like coronary appearances are not generally seen. Recent genetic studies have shown SCAD and FMD share common risk variants including the PHACTR1 locus³¹ but with hints that the effect size may be different between patients presenting primarily with SCAD compared with those with FMD. Common genetic variants associated with general forms of SCAD are also associated with SCAD risk among individuals with FMD,³⁰ suggesting that the two diseases are not one and the same. Likewise, the contribution of causal rare genetic variants may be different. Further work is continuing to understand at a molecular genetic level to what extent SCAD and FMD relate.⁵⁴

Reports of the prevalence of FMD in SCAD have ranged widely depending on modality and completeness of imaging and the permissiveness of the definition used but it is likely this affects between 30% and 50% of SCAD-survivors with the commonest affected vascular beds being the cervical, renal, and ilio-femoral vessels.¹²⁹ On the other side, SCAD may occur in patients with FMD but the incidence appears low in follow-up reports from international FMD registries (2.7% of patients in the US Registry for FMD).¹⁰ So, whilst there is clinical overlap between these conditions, a significant proportion of patients with SCAD have no evidence of FMD when screened using cross-sectional brain-to-pelvis imaging and most patients with FMD will never have SCAD.

Importantly in most SCAD patients, findings of multifocal FMD appear benign and non-progressive in the short to medium term, with, low reported rates of non-coronary major adverse cardiovascular events in SCAD-survivors.^129,130 For example, there have been no reports to date of renovascular hypertension requiring intervention developing in a SCAD-survivor identified on screening to have renal multifocal FMD. Longer follow-up data however, are of considerable importance in this younger population and are currently limited.

It seems clear that SCAD and FMD are closely related pathologies. However, clinical, radiological and genetic data suggest they are overlapping disorders rather than a single disease. Whilst initial progress has been made, further international collaborative research is needed to better understand the pathophysiological relationship and the best clinical management of these important conditions.

12. Visceral phenotype in FMD

Visceral artery FMD (VA FMD) is an uncommon manifestation of FMD, with presentations varying from incidentally discovered to symptomatic or even life-threatening disorder.^3,4,8,19,132

Its prevalence among FMD patients is estimated at 15–20% (range 9–48.7%), based on various cohorts and comprehensiveness of vascular screening.¹³²

In FMD patients enrolled in ARCADIA-POL who underwent a full brain-to-pelvis vascular screening, the prevalence of VA FMD was 13.8%.¹⁹ Most cases of VA FMD were asymptomatic. However, in five patients (15.6%), VA FMD presented as abdominal emergencies e.g. rupture of a hepatic artery aneurysm or acute thrombosis of mesenteric artery.

In this series, VA FMD affected most frequently the coeliac trunk (83.1%). Other arteries were affected less frequently: superior mesenteric (25%), splenic (9.1%), common hepatic (6.3%), and inferior mesenteric (6.3%). Of interest, 42 aneurysms in visceral arteries were found in 20 patients with VA FMD (62.5% of this cohort). Moreover, patients with VA FMD were characterized by lower BMI as compared to patients with only renal FMD, healthy controls and patients with resistant essential hypertension.¹⁹

Furthermore, compared to matched controls, FMD patients were characterized by significantly smaller diameters of all visceral arteries irrespectively of the VA FMD presence.¹⁹ In line with the aforementioned findings of Bruno et al. in brachial arteries, this anatomical difference may represent a novel phenotypic expression of FMD.¹⁹ This observation nevertheless requires confirmation in larger cohorts and other vascular beds.^19,85

13. Pregnancy-related complications

Exogenous or endogenous female hormones have been reported as potential risk factors for FMD^133,134 but no strong evidence of a clear association with FMD has yet been identified. In a small renal histological study, an intense progesterone receptor expression in the nuclei of smooth muscle cells of the renal arteries was detected in six patients with FMD but not in three control subjects, suggesting that progesterone may play a role in the pathogenesis of FMD, but further studies are needed to confirm these findings.¹³⁵

Until recently, data on pregnancy-related complications in patients with FMD were mostly based on the results of a single retrospective study¹³⁶ reporting a very high risk of superimposed pre-eclampsia in a small cohort of patients subsequently revascularized for FMD-related renal artery stenosis. Furthermore, an increased risk of worsening of hypertension status and of vascular complications (aneurysms rupture or arterial dissection) during the peri-partum period has been postulated in FMD patients according to several case reports and expert opinion.^137–140

Very recently, the nature and prevalence of complications occurring in FMD patients during pregnancy or within 3 months postpartum were extensively assessed in the European/Internal FMD Registry.²¹ Out of 237 women with FMD, 40% experienced pregnancy-related complications. Compared with women from the general population,^141–143 patients with FMD had a substantially higher prevalence of gestational hypertension (25% vs. 3.6–9.1%) and preterm birth (20% vs. 5.5–12.2%), while pre-eclampsia was only moderately increased (7.5% vs. 1.6–4%) (Figure 8). When compared to women with primary hypertension,^142,144,145 both prevalence of preterm birth and, especially, of pre-eclampsia was lower in women with FMD (20% vs. 22.7–33.3% and 7.5% vs. 22–25.9%, respectively). Other maternal placental syndromes, such as abruptio placentae and intra-uterine foetal death, as well as severe vascular complications (arterial dissection and/or aneurysm rupture) were rare (Figure 8). Finally, only <5% of women with FMD developed resistant hypertension during pregnancy.

Figure 8.

Pregnancy-related complications in patients subsequently diagnosed with FMD enrolled in the European/International FMD Registry.²¹ Gestational hypertension—25%, 0%, and 9.1%; preterm birth—20.1%, 33.3%, and 12.2%; pre-eclampsia—7.5%, 25.9%, and 4.0%, for patients with FMD, chronic hypertensives, women from the general population, respectively.

Since complicated pregnancies were more frequent in younger patients with renal FMD and a higher rate of renal artery revascularizations,²¹ the higher prevalence of pregnancy-related complications compared to the general population may be due to the severity of renovascular hypertension. Still, other less well-studied factors, such as increased inflammation, endothelial dysfunction, changes in cytokines (i.e. TGF-β1 and -β2), and changes in the VGEF/PlGF balance or female hormones levels may account for this increased risk. Further studies are needed in order to replicate these results and dissect the underlying mechanisms.

14. 2019 FMD consensus

In 2019, the First International Consensus on FMD was published simultaneously in the Journal of Hypertension and Vascular Medicine.^1,2 This document built upon 2014 scientific statements from the USA and Europe, with a writing committee commissioned in 2017 by the Working Group ‘Hypertension and the Kidney’ of the European Society of Hypertension (ESH) and the Society for Vascular Medicine (SVM).^6,146 The goal of the International Consensus document was to create a single, up-to-date expert consensus on adult FMD that would replace the two 2014 documents. In addition to European and American specialists representing the ESH and SVM, the writing committee included experts from other regions of the world and representatives of patient advocacy groups. The document includes 13 consensus-based points covering multiple aspects of FMD with the acknowledgement that there was generally limited level I data available on which to make recommendations. As explained, the intent of the writing committee was that the document, ‘including identification of research priorities, will lead to future high-quality research efforts, additional observational studies, and randomized controlled trials, and that these data will be incorporated into a future international guideline document’. Important clinical consensus points in the document address brain-to-pelvis imaging for patients with FMD, the prescription of antiplatelet therapy for patients with FMD in the absence of contraindication, and the use of a standardized protocol for catheter-based renal angiography and angioplasty that incorporates haemodynamic assessment of translesional renal artery pressure gradients.^1,2 In addition to the consensus points and research priorities (Table 3), the document includes a summary of FMD mimics, comparison of data in the US and European FMD registries, and a review of international patient advocacy efforts. The co-published document has been widely read with more than 22 000 views since its publication.

Table 3.

FMD research priorities from the 2019 International Consensus. Adapted from Refs^1,2

Further studies to extend the number of genes associated with FMD.

Study of familial patterns of FMD, including intermediate phenotypes, and feasibility/utility of different modalities of screening.

To gain a fundamental understanding of the principal molecular/pathologic processes underlying FMD.

To understand both shared and distinct biologic mechanisms of FMD, SCAD, and cervical artery dissection.

Establishment of a tissue bank of pathology specimens to allow for study of the arterial wall properties associated with FMD.

Risk stratification: what are the factors that determine major adverse cardiovascular events in patients with known FMD? Can we separate out ‘benign’ from ‘severe’ phenotypes, including those with increased risk of dissection/aneurysm?

Determination of whether FMD is a progressive disease (within involved vascular beds, extension to other vascular beds) and the factors associated with disease progression.

Determination of the prevalence of FMD in at-risk patient populations (e.g. young and middle-aged patients with stroke and TIA without apparent CV risk factors, women with severe migraine headaches).

Studies to explore the efficacy of antiplatelet therapy vs. none for primary prevention of cardiovascular events among FMD patients; enrolment stratified by cerebrovascular vs. no cerebrovascular involvement.

Studies to demonstrate the effectiveness of haemodynamic-guided angioplasty (i.e. pressure gradient measurement) for renal FMD in terms of clinical outcomes, including cure or control of hypertension.

15. New directions in FMD research

The past decade has seen significant progress in FMD-related research, which has led to improved understandings of the disease’s clinical manifestations, natural history, epidemiology, and genetics. An imaging-based taxonomy has been developed and an International Consensus has been published, which includes guidance on the diagnostic approach and management of patients with FMD.^1,2 Ongoing work described above continues to focus on FMD genetics and proteomics, physiological effects of FMD on cardiovascular structure and function, and novel imaging modalities and blood-based biomarkers that can be used to identify subclinical FMD. As outlined in the International Consensus, there are a number of high priority research areas within our field that can hopefully be addressed in the years to come through ongoing and future international collaborations (Table 3). In addition to these priorities, exploration of non-vascular manifestations of FMD, including those potentially involving the connective tissues and musculoskeletal system, should be undertaken.^52,147 It is also hoped that the next decade will bring the development of multi-centred and potentially international clinical trials to provide comparative effectiveness data to inform the optimal management of patients with FMD and guide future clinical practice guidelines with level I evidence.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Authors’ contributions

Conception and design of the work: A.P., P.D., A.P., and A.J.

Drafting of the work: A.P., P.D., H.L.G., D.A., R.M.B., M.B., J.-B.D., S.K.G., T.G., M.J., J.C.K., M.K., B.L., M.P., M.P., P.V.d.N., D.J.L.V.T., E.W.-C., A.P., and A.J.

Critical revision for important intellectual content: J.W.O., M.A., P.B., d.P.L., and E.T.

Conflict of interest: This paper was handled by deputy editor Karlheinz Peter.

Funding

This work was supported by Grant of The National Institute of Cardiology no. 2.40/III/19, Poland to P.D, E.W.C., A.P and A.J., National Heart, Lung and Blood Institute/National Institutes of Health (R01HL139672, R01HL122684, R01HL086694), Frankel Cardiovascular Center, and the University of Michigan Taubman Institute to S.K.G., by National Institutes of Health (R01HL130423, R01HL135093, R01HL148167-01A1) and New South Wales health grant to J.C.K., by grants from Beat SCAD, the British Heart Foundation (PG/13/96/3060) and the NIHR Leicester Biomedical Research Centre to DA.