Article Text

Download PDFPDF

Disease severity in children and adolescents with familial Mediterranean fever: a comparative study to explore environmental effects on a monogenic disease
  1. S Ozen1,
  2. N Aktay1,
  3. E Lainka2,
  4. A Duzova1,
  5. A Bakkaloglu1,
  6. T Kallinich3
  1. 1
    Department of Pediatrics, Hacettepe University, Ankara, Turkey
  2. 2
    Center for Pediatric, Rheumatology, University of Duisburg-Essen, Germany
  3. 3
    Pediatric Pneumology and Immunology, Charité University Hospital Berlin, Berlin, Germany
  1. Dr Seza Ozen, Professor, Department of Paediatrics, Hacettepe University Faculty of Medicine, 06100 Ankara, Turkey; sezaozen{at}


Background: Worldwide, familial Mediterranean fever (FMF) is the most common autoinflammatory disease. It has been suggested that environmental factors affect the phenotype as some patients do not develop the complication of secondary amyloidosis.

Objective: To analyse whether disease severity in Turkish children with FMF, living in Turkey and Germany is different.

Patients and methods: A total of 55 Turkish children living in Turkey were compared with 45 Turkish children born and raised in Germany. Mean age among the group from Turkey and Germany was 42.2 and 44.29 months, respectively. M694V was the leading mutation in both groups. The severity scores were compared with two scoring systems, modified according to published paediatric data for dosage.

Results: There was no significant difference between the mean C-reactive protein and erythrocyte sedimentation rate levels of the two groups. According to the modified Sheba Center score, 78.2% of patients from the group living in Turkey had a severe course compared with 34.1% from the group living in Germany. The modified score of Pras et al also showed more severe disease in the patients from Turkey. The difference between the two groups for both scoring systems were significant (both p<0.05).

Conclusions: We believe the modified scores that we introduce can be widely used for children. Our results suggest that the environment affects the phenotype of a monogenic disease of the innate inflammatory pathway.

View Full Text

Statistics from

Defining the gene and the pathogenesis of familial Mediterranean fever (FMF) has opened a new chapter in medicine: the autoinflammatory diseases. Owing to an inborn error in the innate immune system FMF is referred to this group of autoinflammatory disorders. These diseases are often monogenic and are characterised by defects in the inflammatory pathway of interleukin 1 processing.1 2 The gene associated with FMF is the MEFV (Mediterranean fever gene) gene and encodes for a protein called pyrin. FMF is characterised by recurrent periodic fever episodes and serositis along with an increased acute inflammatory response.13 It is the most common autoinflammatory disease and has prevalences as high as 1/1000–1/250 among Jews, Turks, Armenians and Arabs.4 5 FMF has become widespread all over the world with the immigrations of the twentieth century.

It has been long known that environmental factors affect the phenotype as patients in the USA develop the complication of secondary amyloidosis with a much lower incidence.6 As mutations in the protein coded by the MEFV gene are responsible for a disease of innate immunity it may be speculated that the environment may be affecting the phenotype and disease expression. To substantiate this hypothesis we compared the modified severity scores of Turkish patients with FMF who have spent their childhood years in Turkey and Germany.


A total of 55 Turkish children born and raised in Turkey were compared with 45 Turkish children born and raised in Germany. All were under the age of 18 years.

We modified the scoring criteria developed by Sheba Medical Center7 by the integration of recommended age-related doses (table 1). This dose schedule has been previously published by us.8 Additionally, disease severity was determined by the use of the scoring system developed by Pras et al and modified by us (table 2).9 We removed the age factor, which was redundant for children.

Table 1 Familial Mediterranean fever-SS2 severity score by the Sheba Medical Center modified for children (adapted from Mor et al7)
Table 2 Severity score by Pras et al modified for children

All patients from the centre in Turkey and Germany were prescribed colchicine according to the suggested dosage schedule.8 No side-effects were reported.


The results were analysed using the Social Package for Statistical Sciences (SPSS) 11.0 and expressed as median (minimum–maximum) for data not showing normal distribution and as mean (SD) for data showing normal distribution. χ2 test was used for comparing categorical variables and Student t test was used for comparing numerical variables.


The mean age among the group from Turkey and Germany was 42.2 months (range 2–120) and 44.29 months (range 3–178), respectively. M694V was the leading mutation in both groups. The gender and age distribution was comparable in the two groups (table 3).

Table 3 Demographic features, acute phase reactants and severity scores in patients with familial Mediterranean fever

There was no significant difference between the last visit mean CRP and ESR levels of the group from Turkey (mean CRP 0.83 mg/dl, mean ESR 16.9 mm/h) and Germany (mean CRP 0.5 mg/dl, and mean ESR 16.2 mm/h) (table 3). There was no correlation between the disease severity defined by the different scoring systems and the acute phase reactants.

According to the modified severity score suggested by the Sheba Medical Center,7 the median scores of disease severity was 4 in children living in Turkey as compared with 2 in those living in Germany. When patients living in Turkey were assessed with this score 10.9%, 10.9% and 78.2% had mild, intermediate and severe disease respectively as compared with patients living in Germany with 27.3%, 38.6% and 34.1% having mild, intermediate and severe disease, respectively; 78.2% of patients from the group living in Turkey had a severe course compared with 34.1% from the group living in Germany. The ratio of patients with a severe score according to criteria suggested by Pras et al,8 were also higher in the patients born and grown up in Turkey (34.5%) compared with patients living in Germany (15.4%). The difference between the two groups for both scoring systems were statistically significant.


Two scoring systems have been developed to assess disease severity in FMF.7 9 Both have been developed for adults and the one suggested by Pras et al9 mainly adjusts the severity before the introduction of colchicine treatment. We have modified these scores by including the age-adjusted dose and the current parameters for subclinical inflammation. These two additional parameters reflect the current disease severity under the already introduced colchicine therapy. Furthermore, the age factor was removed as it was redundant for children. This was the first time these scores were used in children. We believe that the modified scoring system may be used widely in paediatric practice.

When testing the reliance of the acute phase reactants, e.g. CRP and ESR, no correlation with the scoring system was found. This observation was noteworthy as clinicians often use the acute phase reactants to assess disease severity per se. As serum amyloid A levels were not consistently measured (in the German centres) a statistically valid analysis could not be performed. Future studies will show whether other parameters of subclinical inflammation are required in the assessment of severity among patients with FMF. It is probable that the severity of disease is different in patients before colchicine is started and the difference becomes less obvious after treatment (unpublished observations). This again necessitates further studies.

It has been shown that the country where the patients originate from is a risk factor for the development of amyloidosis. Originally, this was demonstrated in a cohort of Armenian patients living in the USA; although the development of amyloidosis is frequently observed in Armenia this complication did not occur in a cohort of Armenian patients living in California.6 Similar conclusions can be drawn from a recently published multicentre study.10 Thus, it is implicated, that not only genetic but also environmental factors are predictors for the development of secondary amyloidosis.

Epigenetics is the interactions of genes with their environment that brings the phenotype into being. Epigenetic mechanisms such as DNA methylation have not been specifically studied in FMF. However, various polymorphisms in other genes within the inflammatory pathway have been previously shown to affect the development of amyloidosis, which may be regarded as affecting the phenotypic expression of the disease. Secondary amyloidosis is the most severe complication of the disease and occurs only in patients with the most intense expression of inflammation. A number of genetic factors such as polymorphisms in the gene coding for serum amyloid A, have been shown to be associated with the development of secondary amyloidosis.11 12 We have previously shown that the Arg753Gln TLR2 polymorphism affects the disease severity as well.13 We hypothesised that this TLR2 polymorphism may be an important link defining the severity of the patients in this area where the common microbes are more frequently encountered than in the Western countries.13

In this study we show objective parameters suggesting that the country of residence influences the phenotype of FMF in children and adolescents. The children in this study are of the same ethnic origin, have similar genotypes and probably similar polymorphisms. The only difference is the environment, where they have spent their early years of childhood, a time when they contact with the common microbes. Pyrin is a component of the inflammasome affecting the interleukin 1 pathway. The interleukin 1 pathway interacts with many inflammatory pathways of pathogen-associated molecular patterns as well. Thus, the exogenous microbial triggers may affect the final disease expression. There is no study comparing the microbial flora of eastern Mediterranean and the European countries. However, the different rates of acute rheumatic fever and post-streptococcal glomerulonephritis suggest the occurrence of diverse bacterial milieus. This view is also supported by the observation that the rate of allergic sensitisation in children originated from Turkey, and living in Berlin is correlated to the grade of cultural adaptation. Thus, environmental circumstances, eg the exposure towards microbial agents, strongly contribute to the phenotypical expression of this multigenetically inherited disorder,14 leading to different disease prevalences in Turkey and Germany. On the other hand, diseases associated with superantigens such as Kawasaki disease is less frequent in Turkey.15 The exact reason for the different phenotypic expression of FMF remains unclear. Besides differences in the microbial environment other causes such as the feeding pattern or the living conditions has to be considered. To delineate the impact of these conditions on the phenotypical expression of FMF, a prospective study performed in different countries is warranted.

This is a pilot study that may need to be repeated in large studies in different geographic areas. However, this is the first comparative cross-sectional study showing that the environment affects the phenotype of a monogenic disease—FMF. The difference we show between the two groups in a disease of innate immunity, may also support the hygiene hypothesis.


View Abstract


  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.