Marfan syndrome (MFS) is an autosomal dominant genetic connective tissue disorder diagnosed by the Ghent criteria.1 To obtain proper counselling and care of individuals with MFS, a precise diagnosis is mandatory, prolonging life expectancy and reducing disability.2 The prevalence of MFS is disputed. Two MFS prevalence studies reports 6.81/100.0003 and 4.6/100.000 inhabitants, respectively,4 whereas the usually cited prevalence is 1-2/10.000 inhabitants.5

The Ghent criteria describe a number of variably well-defined symptoms and signs. They are categorized as ‘major criteria’, ‘manifestations’ (in the skeletal system) and ‘minor criteria’, which alone or combined indicate the ‘involvement of an organ system’ or the ‘fulfilment of a major criterion’. Fulfilment of at least two major criteria in two different organ systems and the involvement of a third organ system are required to give the diagnosis.

Although FBN1 is the only gene included in the Ghent nosology,1 FBN1 mutations are also found in individuals not fulfilling the Ghent criteria,6, 7 and mutations in TGFBR1 and TGFBR2 have recently been reported in individuals fulfilling the Ghent criteria.8, 9

The resources needed and the logistic challenge in the examination of all organ systems may explain why clinical papers do not commonly present complete data for all variables. Furthermore, some may consider it appropriate to stop the investigations when data sufficient for a diagnosis of MFS have been acquired. This is illustrated by the fact that less than 30% of the individuals were investigated for dural ectasia and protrusio acetabuli in recent collective reviews on the basis of the FBN1 mutation database, presenting genotype–phenotype correlations.10, 11 Moreover, various prevalences of ‘major criteria’ and ‘organ involvement’ are presented. The various results might in part be owing to the features being age dependent,12 as most studies present results for children or mixed age groups whereas reports for adults are rare.13, 14

The need for sensitivity and specificity of the ‘major criteria’ and ‘organ involvement’ has been pointed out.1, 15 However, neither prevalences nor sensitivity and specificity are known.

Thus, the purpose of this study was to explore the phenotype, the prevalence of each ‘major criterion’ and ‘organ involvement’ through a prospective and complete investigation of all features of the Ghent criteria in an adult cohort with a proven diagnosis of MFS.

Patients and methods

This study was approved by the regional ethics committee. In 2002, Norway had 4.5 million inhabitants, the majority being Caucasians. By October 2002, 179 individuals had reported themselves to the National Resource Centre for MFS (TRS) as having MFS; 134 of them were 18 years of age or older and all were citizens of Norway.

The participants in the Norwegian Marfan Syndrome Study were recruited either by an invitation letter sent to the 134 individuals above 18 years of age in the TRS database, who had registered as having MFS, where by 80 persons signed in for participation, or by an advertisement in the journal of the Norwegian Association for MFS asking for persons who have been diagnosed with MFS, or through invitations distributed in the Department of Thoracic and Cardiovascular Surgery, Rikshospitalet University Hospital to patients suspected for MFS, thereby recruiting an additional 19 persons, although the number of persons informed about the study is unknown. A total of 109 individuals thus gave their informed consent to participate, but one died before the study started, one was not able to attend while living abroad and two participants withdrew, one before and one during the study.

Consequently, the study population consisted of 105 individuals, 67 women (64%) of median age 42 years (range 20–69 years) and 38 men (36%) of median age 33 years (range 19–62 years).

Ninety of the 105 had earlier been given a diagnosis of MFS; 15 persons entered the study because of suspicion of MFS. All participants were Caucasian.

The 105 individuals represented 66 families. Forty-five individuals were the only representatives of their family, whereas 60 individuals were from 21 families.

To obtain valid results, the same group of investigators, using the same methods for all patients, carried out all examinations. All participants were assessed for all Ghent criteria.1 The methods of assessment of clinical and radiological examinations and cutoff limits are presented in Table 1.16, 17, 18 Details of the mutation analysis have been published.19, 20

Table 1 Diagnostic criteria for Marfan syndrome (MFS) according to the Ghent nosology

As recommended by the authors of the Ghent criteria, the serum concentration of homocysteine was measured in all 105 individuals to exclude homocystinuria.1

When all data were present, the status for each individual was assessed with regard to each organ system (major criteria fulfilled and organ systems involved). Adding up, the number of major criteria fulfilled and organ systems involved was registered, and the individual was classified as fulfilling the Ghent criteria or not.


All data were stored in a customized database (SPSS for Windows version 13). Descriptive statistics are reported as medians (ranges). Owing to the low number of individuals, results are reported as fractions and percentages.


Of the 105 individuals, 87 persons from 57 families fulfilled the Ghent criteria (83%), whereas 18 individuals (17%) from 9 families did not. Of those individuals fulfilling the Ghent criteria, 77 had earlier been diagnosed with MFS, whereas MFS was suspected in 10 individuals. The summary and interpretation of all examinations and the prevalences of fulfilled major criteria and involved organ systems are presented in Table 2 and Figure 1.

Table 2 Prevalence of the individual features.
Figure 1
figure 1

Number of persons fulfilling major criteria and having organ systems involved. Over all, 105 persons with given or suspected diagnosis of Marfan syndrome. Order as in the Ghent paper.

The prevalence of the major criteria in the other four organ systems did not show significant differences in accordance to gender.

The 87 persons fulfilled the Ghent criteria through 56 different combinations of major criteria and organ involvement; the highest number of individuals having the same combination of criteria was 5. Ten persons fulfilled all five possible major criteria, three of them also having lungs and skin and integument involved.

Of the 87 fulfilling the diagnostic criteria, 76 fulfilled the Ghent criteria on the basis of the individual examinations alone, without considering family history or mutation analyses.

Another nine individuals did fulfil one major organ criterion and had a first-degree relative independently fulfilling the Ghent criteria (positive family history) as the second major criterion and a third organ system involved.

Two individuals, both being the first in their family to be diagnosed with MFS, fulfilled one major organ criterion, had an FBN1 mutation as their second major criterion and had a third organ system involved. These two persons were the only ones to be dependent on mutation analysis to be diagnosed with MFS.

Of the 87 persons fulfilling the Ghent criteria, 14 were dependent on the dural criterion for the diagnosis.

Of the 87 individuals fulfilling the diagnostic criteria, an FBN1 mutation was found in 73 individuals, representing 44 families.19, 20 No FBN1 mutations were found in individuals not fulfilling the Ghent criteria.

Of the 18 individuals not fulfilling the diagnostic criteria, 13 had been diagnosed with MFS earlier, including a person with a highly elevated serum concentration of homocysteine (320μmol/l), one person in whom a mutation in the Collagen 3 gene (COL3A1) was found and five persons from one family in whom a TGFBR2 mutation was found.

The prevalence of the features in the Ghent criteria among the 87 individuals fulfilling the diagnostic criteria in our study is presented in Table 3 for comparison with prevalences from other studies reporting the prevalence of two or more major criteria.11, 21, 22, 23, 24, 25, 26, 27, 28, 29

Table 3 Prevalence (%) of major criteria and affected organ systems in the Ghent criteria in papers presenting 2 organ systems

The prevalence of having at least two major criteria fulfilled in our study is presented in Figure 2.

Figure 2
figure 2

Number of persons fulfilling two major criteria, organized after prevalence. N=105.


This study confirms the need for the complete Ghent criteria to identify all patients with MFS, although the majority of them might be identified by combined assessment of dura and family history, supplemented with DNA analysis in family-negative cases.

Using the full version of the Ghent criteria on 105 individuals refuted the MFS diagnosis in 13 of 90 patients who had been diagnosed earlier, and verified the diagnosis in 10 of 15 persons with suspected MFS. The 56 different ways of fulfilling the Ghent criteria in this cohort illustrate the need for the complete set of criteria and the problems with differential diagnosis versus other genetic connective tissue disorders. Our cohort represents persons with given or suspected diagnosis of MFS; the primary clinical result of our investigations is the separation of individuals fulfilling the Ghent criteria from individuals suspected for MFS, not fulfilling the Ghent criteria.

The Ghent criteria require at least two major criteria to be fulfilled. The major criteria in five organ systems have equal value and validity in the diagnostic process. Even so, the concept of severity of Marfan phenotypes has been presented.30, 31, 32 ‘Severe Marfan syndrome’ often seems to be understood as ‘severe cardiovascular pathology’, irrespective of age, not as a person fulfilling many major criteria. The term ‘severity’ in accordance with the MFS has yet to be defined.

In spite of the great diversity of combinations, a combination of the dural major criterion and the presence of genetic major (positive family history and/or mutation in FBN1) criteria could identify 69 out of 87 (79%) affected individuals with MFS in our cohort.

The lack of investigations on the dura in many other studies and the problems resulting from this have been pointed out by Adès et al33

Dural ectasia was the most frequent major criterion present in individuals fulfilling the Ghent criteria (91%) comparable with the prevalences found in other studies.15, 21, 34 It is noteworthy that 16 % of the individuals fulfilling Ghent were dependent on this major criterion. If dura had not been investigated, they would not have been given the diagnosis.

In our study, a positive family history (64%) was found less often than expected. This is probably because of the high number of probands in our study and because investigations of family members were not carried out for relatives who had not signed in for the study; the participants were asked about their relatives having obvious major manifestations as known aortic dilatation or operation, known ectopic lenses or lenses removed and known mutations; see subtext to Table 1.10

In spite of sequencing all coding parts of FBN1 and searching for large deletions and duplications, the presence of FBN1 mutations (44 probands out of 57 probands fulfilling the Gent criteria, 77 %), is lower in our study than in that commonly reported among individuals fulfilling the Ghent criteria.2, 35 All patients with a mutation in FBN1 did fulfil the Ghent criteria; thus, no other ‘type-1 fibrillinopathies’ were found.7 Investigating dura in all cases may be the reason for this discrepancy.

The prevalence of ascending aortic disease among individuals fulfilling the Ghent criteria in our study (53%) (Table 2), more often found in men than in women, is among the lowest reported. We have not found papers reporting different prevalences of major aortic pathology depending on gender. However, families have been observed in which the men are more likely to have earlier onset aortic enlargement, more rapid enlargement and earlier dissection than the women in the family.36 One might speculate whether our results represent a true difference between genders or differences between recruiting men and women. In our clinical work, it seems that a man needs a serious organ affection as a trigger for searching for help in their coping process. The prevalence of mitral valve prolapse (MVP) was also low compared with that in other studies, which, however, presents cohorts of children or mixed age groups.37, 38, 39, 29 We did not encounter any patient with a dilated pulmonary trunk who was below the age of 40 years, using the aortic normograms as suggested in the Ghent paper. Using Nollen's upper limit of normality for the pulmonary trunk, 34.8 mm, 13 out of 87 persons fulfilling Ghent had enlarged pulmonary trunk (median diameter 30 mm; range 23–38 mm).40

As most studies that were found have emerged from specialized 4th level cardiovascular centres serving individuals with severe aortic or cardiovascular disease, patient selection may explain the high prevalence of cardiovascular pathology in other reports.

The acronym ‘MASS phenotype’ has been suggested to emphasize the involvement of the mitral valve, aorta, skeleton and skin in persons who could not be classified in accordance to the Berlin nosology.41 In our study, no patient fitted into the ‘MASS phenotype.’

Our cohort represents persons recruited through all medical specialities and through the patient organization. In addition, the social security system in Norway does prevent economical constraints for diagnosis and participation in the study. Our study cohort may therefore be more representative for an adult population with MFS of different phenotypes.

Most studies of cardiovascular pathology in MFS have been carried out on groups with low mean age; therefore the natural history of the aging Marfan aorta is not well known. In our adult cohort, we expect the features of MFS, including cardiovascular manifestations, to be present, and it is unlikely that typical features would disappear during life. Although a further development of aortic pathology in our cohort in the next few decades cannot be excluded, our study indicates that dilatation and dissection of the ascending aorta and of MVP may be found more seldom in adults than in cohorts including Marfan children and adolescents. This underlines the more severe consequences of manifest cardiovascular disorders in the lower age groups, sometimes resulting in early cardiovascular death, as illustrated by Gray et al.42 Thus, a patient selection on the basis of the natural history of the disease may have taken place. However, if the low prevalence of cardiovascular pathology is representative for adult persons with MFS, it will influence the overall clinical outcome and, consequently, the estimated individual prognosis when counselling adult persons with MFS.

The prevalence of lens subluxation or luxation in our study, 62%, is the average of what is reported,43, 4 whereas the prevalence of fulfilling the major criteria for the skeletal system was 38%; other reports show large variation.23, 24 The involvement of skin and integument is comparable with the highest prevalence published.23, 27, 44

Spontaneous pneumothorax has been rare in our cohort, whereas blebs in the lungs were found more often.45 This can partially be explained by different imaging techniques.

The patient with two major criteria, but not fulfilling the Ghent criteria, had subluxation of one lens and a mother independently fulfilling the Ghent criteria. The mother had the family mutation, whereas the patient did not.

Family history, aortic disease, loose lenses or Marfanoid habitus usually raises the suspicion of MFS. Our results indicate that at the time of suspecting the MFS in an adult, an early investigation using MRI of the lumbosacral spine to detect dural ectasia is appropriate. When MFS is confirmed, the individual should be referred to a interdisciplinary ‘Marfan centre’ for individual counselling, follow-up and care.2 Even if close to 80% of our cohort could have settled the diagnosis by dura and family/genetic major criteria, a complete examination of all organ systems in the Ghent criteria should be carried out.

The broad variety of ways of fulfilling the Ghent criteria illustrates the need for individual counselling on the basis of the individuals’ own signs and findings in all organ systems.

Longitudinal studies of groups of persons fulfilling the Ghent criteria might indicate whether some variants may have more benign natural history than do others. Until then, all persons fulfilling the Ghent criteria should have ophthalmological and aortic controls for the rest of their lives.

A strength of our study is the defined population from which the participants are recruited (the Norwegian population of 4.5 million inhabitants) and that the same group of investigators using the same methods carried out all the examinations with close to no missing data. The complete investigation in all cases may have increased the number of individuals fulfilling the Ghent criteria through combinations of features not depending on the cardiovascular system. The relatively small size of our cohort might possibly influence our results. However, most commonly, current papers provide incomplete clinical information as compared with the extensive and complete investigations performed in this study.

A limitation is that our study population was skewed for gender, women being in surplus. This reflects the representation of gender among individuals over 18 years of age with MFS, who have registered themselves at the National Resource Centre for Rare Disorders, TRS.

In conclusion, using the complete list of features in the Ghent criteria, out of 105 adult individuals with given or suspected diagnosis of MFS, 87 fulfilled the Ghent criteria in 56 different variants. The large number of variants shows the need for using the complete set of features in the Ghent criteria. Dural ectasia is the major criterion most often present in persons fulfilling the Ghent criteria (91%), followed by FBN1 mutations (84%), positive family history (64%) and ectopic lenses (62%). Dilatation or dissection of the ascending aorta was only found in 53% of Ghent-positive persons (30 of them operated), one of the lowest prevalences published. The gender difference observed in the prevalence of major aortic pathology may be real or may represent a selection bias. Having dural ectasia as well as fulfilling the major genetic criterion (positive family history and/or FBN1 mutation) was the combination of fulfilling two major criteria most often found, with 79% of persons fulfilling the Ghent criteria, indicating an early investigation of those systems when suspicion of MFS has been raised in adults. The low prevalence of cardiovascular pathology might indicate better future prospects in an adult population than those traditionally considered. More studies are needed describing the prevalence of all Ghent features, thereby making it possible to calculate the sensitivity and specificity for each feature.