Familial Mediterranean Fever

A number sign (#) is used with this entry because familial Mediterranean fever is caused by mutation in the pyrin gene (MEFV; 608107).

Description

Familial Mediterranean fever (FMF) is an autosomal recessive disorder characterized by recurrent attacks of fever and inflammation in the peritoneum, synovium, or pleura, accompanied by pain. Amyloidosis with renal failure is a complication and may develop without overt crises (French FMF Consortium, 1997).

See also autosomal dominant FMF (134610), which is caused by heterozygous mutation in the MEFV gene.

Clinical Features

Siegal (1945) reported 'benign paroxysmal peritonitis.' Under the term 'periodic peritonitis,' Reimann et al. (1954) described 72 cases from Lebanon, most of them Armenian. In 1 remarkable family, survivors of the siege of Musa Dagh, 20 affected persons occurred in 5 generations, with 3 instances of skips in the pedigree. The high gene frequency and small breeding group could account for the findings as representing pseudodominant inheritance. Sokmen (1959) noted that many cases of FMF in Turkey were observed in persons without known Armenian ancestry. This condition was called 'familial paroxysmal polyserositis' by Siegal (1964), who was the first to delineate the disorder clearly in the United States and who observed rather numerous cases in Ashkenazim. Barakat et al. (1986) referred to FMF as 'recurrent hereditary polyserositis.'

Armenian (1983) compared the characteristics of 79 patients seen for the first time in a special clinic for FMF in its first 16 months of operation with the characteristics of 79 patients who presented during the last 6 years of its operation. The patients studied during the first period had a more severe form of the disease with multiple clinical manifestations such as proteinuria and amyloidosis. There were more males and more patients with a positive family history in the earlier group. Armenian (1983) emphasized the importance of a population base and 'enrollment bias'--differences in referral pattern, in case selection, and in the sources of data--in accounting for significant variation in the frequency of various clinical manifestations in published series of FMF cases.

In a study of FMF in Kuwait, Barakat et al. (1986) reported on an 11-year experience with 175 Arab patients. The most common manifestation was peritonitis (94%), followed by arthritis (34%) and pleurisy (32%). Amyloidosis, rashes, hepatosplenomegaly, and lymphadenopathy were rare.

Mollaret (1944) described a syndrome of benign, recurrent meningitis with a characteristic spinal fluid picture: pleocytosis of mixed cellular type including endothelial cells (Mollaret cells) during the attack, in the absence of any positive agent. These attacks were separated by symptom-free periods lasting from days to years. Recovery was complete, with no neurologic deficit. Barakat et al. (1988) concluded that Mollaret meningitis is a feature of FMF. Schwabe and Monroe (1988) described a 32-year-old non-Ashkenazi Jewish man in whom meningitis first developed in Morocco at the age of 1 year. Between the ages of 26 and 31 years, he had 6 attacks of fever, frontal headache, nausea, and stiff neck, with positive Kernig and Brudzinski signs. In Saudi Arabia, Majeed and Barakat (1989) found that 48 of 88 affected children had onset before the age of 5 years.

In a 60-year-old Arab man with FMF, Agmon et al. (1984) observed selective amyloid involvement of the zona glomerulosa of the adrenal cortex resulting in isolated hypoaldosteronism. As a rule, the glomerulosa is spared in adrenal amyloidosis of FMF. Knecht et al. (1985) found abnormally high levels of serum amyloid A (SAA; 104750) in attack-free intervals and very high levels at the onset of attacks. Although there is a striking difference in the frequency of amyloid nephropathy in different ethnic groups with FMF, the elevation of SAA during and between attacks is the same, and interethnic marriages produce affected progeny (Pras et al., 1982).

Rawashdeh and Majeed (1996) reviewed findings from the FMF pediatric patient population in northern Jordan (all children were of Jordanian, Palestinian, or Syrian origin). The 192 patients first presented between the age of 4 months and 16 years. The mean delay in diagnosis was 3.7 years and was increased for children who presented before the age of 2 years. Abdominal pain was the most common presenting symptom and occurred in 51%, while arthritis and pleuritis occurred in 26% and 23%, respectively. The investigators noted that family history was positive in 62% of the children, which was not surprising in this autosomal recessive disorder given the 64% consanguinity rate in northern Jordan. Minimum prevalence was given as 1 in 2,600 with an estimated gene frequency in the childhood population of 1 in 50 (calculated on the numbers of diagnosed patients). The authors warned that the frequency of FMF among Jordanian Arab children was greater than previously estimated.

Eshel et al. (1988) suggested that acute, unilateral, short-term orchitis is a feature of FMF; they observed 20 episodes in 13 patients. Ozyilkan et al. (1994) suggested that bronchial asthma is abnormally infrequent in FMF patients. This was the case not only in patients on colchicine because an absence of a history of asthma before the diagnosis of FMF was also found.

In Ankara, Turkey, Saatci et al. (1997) analyzed 425 FMF patients without and 180 with amyloidosis. Of the latter group, 103 had amyloidosis type I and 57 had amyloidosis type II. Type I amyloidosis was defined as amyloidosis developing subsequent to clinical features of FMF, whereas type II was defined as amyloidosis developing as the initial manifestation. The male-to-female ratio was higher in the amyloidosis population (111 to 69) than it was in the FMF population without amyloidosis (225 to 200) (P = 0.048). The consanguinity rate was the same in the 2 groups. A family history of amyloidosis was significantly more frequent in the amyloidosis group (P = 0.0001). The combination of positive family history of amyloidosis and consanguinity increased the risk of amyloidosis more than 6 fold. The 5-year chronic renal failure-free survival was 43.1% and 18.7% in type I and type II amyloidosis, respectively. Saatci et al. (1997) found 10 cases of Henoch-Schonlein purpura and 9 of polyarteritis nodosa among their patients. The significance of the association between FMF and vasculitis was unclear. Among 435 FMF patients treated with colchicine, only 10 (2.3%) developed amyloidosis, thus confirming that this drug protects patients from the complication.

In a retrospective study of 4,000 FMF patients, using a computer chart review, Kees et al. (1997) found that during a 20-year period, one or more episodes of pericarditis were recorded in 27 patients. Each patient experienced 1 to 3 pericarditis attacks, lasting a mean of 4.2 days, accompanied by elevated temperature and symptoms of FMF attack at another site. The pericarditis resolved spontaneously and left no sequelae.

Commenting on a paper by Yazigi and Abou-Charaf (1998), Tutar et al. (1999) noted that recurrent pericarditis can be the initial sole manifestation of FMF and suggested that mutation analysis for FMF be considered in patients with idiopathic recurrent pericarditis, especially if they are of Mediterranean origin.

Pras et al. (1998) compared the clinical features of FMF in North African Jews and Iraqi Jews, the 2 largest population groups suffering from the disease in Israel. North African Jews were found to have more severe disease manifested by earlier age of onset, increase in frequency and severity of joint involvement, higher incidence of erysipelas-like erythema, and higher dose of colchicine required to control symptoms.

Cattan et al. (2000) studied the incidence of inflammatory bowel disease (IBD; 266600) in non-Ashkenazi Jewish patients with FMF. Cattan et al. (2000) estimated a prevalence of at least 3 per 300 (or 3 per 173 if the calculation is done through probands) in non-Ashkenazi Jews with FMF. They postulated that the inflammatory processes of FMF and IBD are additive, resulting in increased severity of disease in the new patients.

Tamir et al. (1999) characterized a late-onset form of FMF with distinct clinical, demographic, and molecular genetic characteristics. This subset of patients experienced their first FMF attack at age 40 years or later. The comparison group consisted of 40 consecutive FMF patients who arrived at the FMF clinic for their regular follow-up visit and were 40 years of age or older at the time of examination. Only 20 of 4,000 (0.5%) patients had late-onset FMF. These patients were mostly men, of non-North African origin (P less than 0.05 compared to controls). All had abdominal attacks and in most these were the only manifestation of their disease (P less than 0.001). None had chronic or prolonged manifestations of FMF such as amyloidosis, chronic arthritis, or protracted myalgia (P less than 0.001). The response to treatment was good despite using low colchicine dosage (P less than 0.05). The overall severity score indicated a mild disease (P less than 0.001). Mutation analysis revealed absence of M694V homozygosity (P less than 0.01), compared to the regular FMF population. Tamir et al. (1999) concluded that the onset of FMF at a late age defines a milder form of disease with typical clinical, demographic, and molecular genetic characteristics.

Other Features

Schwabe and Lehman (1984) reviewed the search for the basic defect in this disorder. Normal peritoneal fluid contains an inhibitor of neutrophil chemotaxis that acts by antagonizing the complement-derived chemotactic anaphylatoxin C5a (113995). The inhibitor resembles a substance found in synovial fluids and is a protein with molecular weight 40,000. Matzner and Brzezinski (1984) found that this inhibitory activity was less than 10% of normal in peritoneal fluid from FMF patients. Inadequate suppression of the inflammatory response to C5a that is released accidentally may be responsible for the inappropriate inflammatory reactions of FMF. Colchicine may prevent attacks by suppressing neutrophil motility and blocking their mobilization to sites of C5a release. Matzner et al. (1984) found a decrease of the C5a inhibitory activity in synovial fluid from patients with FMF. Patients with other forms of inflammatory arthritis and osteoarthritis had normal levels. Ayesh et al. (1990) further characterized the 40-kD inhibitor protein and showed that it is a serine protease. Documentation of a carrier state of reduced C5a-inhibitor activity in unaffected obligatory heterozygotes such as parents or offspring would help establish the inhibitor as the site of the primary defect. The patients studied by Matzner and Brzezinski (1984) were Sephardim; comparable studies in Armenians with FMF would be of interest because of the much lower frequency of amyloidosis with FMF in this ethnic group (Schwabe and Peters, 1974).

Inheritance

Rogers et al. (1989) presented evidence indicating autosomal recessive inheritance of FMF in Armenians. Using extended pedigree data, they calculated an FMF gene frequency of 0.073 and a carrier rate of 1 in 7, which is about 4 times the frequency in non-Ashkenazi Jews. Four of 64 families had 1 parent affected as well as the proband; the high gene frequency could explain this phenomenon. The male/female ratio of 1.7 found in non-Ashkenazi Jews indicates reduced penetrance in females and probably also obtains in Armenians.

Shohat et al. (1992) found concordance for FMF in all 10 monozygotic twin pairs and only 3 of 11 dizygotic twin pairs. However, variability in the clinical manifestations and degree of severity were noted within twin pairs, supporting the contention that the lower than expected incidence of FMF observed in segregation analysis is due to genetically affected but clinically undiagnosed patients.

Yuval et al. (1995) found 77 families, with 240 FMF patients, in which the disorder affected more than one generation. In 75 of these families, the occurrence of FMF in more than one generation was found to be consistent with a recessive mode of inheritance due to a high gene frequency and consanguinity of the parents. In 2 families, however, one of Ashkenazi and the other of Georgian Iraqi origin, in which FMF occurred in 4 consecutive generations, the transmission could be explained only by autosomal dominant inheritance. Aksentijevich et al. (1999) found that the inheritance in the Ashkenazi family reported by Yuval et al. (1995) as an unusual instance of dominantly inherited FMF was in fact recessive inheritance of the E148Q mutation (607108.0005) which has a high frequency and reduced penetrance among Ashkenazi Jews.

Diagnosis

By means of a placebo-controlled, double-blind, crossover study, Barakat et al. (1984) demonstrated that intravenous infusion of 10 mg of metaraminol bitartrate ('Aramine') in 500 ml normal saline over a period of 3 to 4 hrs was followed by a typical attack of FMF in all of 21 persons with the disease and in none of 21 control subjects. The induced attacks were milder and of shorter duration than the spontaneous ones. The metaraminol-induced attacks could be prevented with colchicine. In connection with the metaraminol provocative test for the diagnosis of FMF (Cattan et al., 1984), Barakat et al. (1984) suggested that abdominal tenderness should be included as a feature indicating positive test.

Fischel-Ghodsian et al. (1993) identified 2 flanking markers and microsatellite markers on chromosome 16p that allowed preclinical diagnosis in most pedigrees with affected persons. Dupont et al. (1997) proposed the use of a set of 7 microsatellite markers for diagnosis and heterozygote analysis: D16S283 and D16S3124, telomeric of the FMF disease gene; D16S3070, D16S3082, and D16S3275, which showed no recombination with disease; and D16S2622 and D16S3027, centromeric of the FMF locus. They found this set of markers to be informative in 100% of previously diagnosed non-Ashkenazi Jewish patients. In addition, 73% of patients from this population were homozygous for the 3-3-9 or 3-3-18 haplotype at D16S3070, D16S3082, and D16S3275. Dupont et al. (1997) suggested that these markers could be used for diagnosis of sporadic cases in this population, although absence of homozygosity would not exclude the diagnosis.

Brik et al. (2001) looked for the presence of the 5 most common FMF mutations (M694V, V726A, M694I, M680I, E148Q) in 59 Sephardic Jewish and Israeli Arab children who had been diagnosed with functional abdominal pain. Eight (34.7%) of the Arab children and 4 (11%) of the Sephardic Jewish children were genetically diagnosed as having FMF. Ten children were heterozygous for 1 mutation. The authors suggested that a higher percentage of Arab children were genetically diagnosed since FMF in Arabs is a milder disease and more likely to be undiagnosed clinically. They noted that testing may spare patients from a fruitless workup and allow institution of colchicine therapy to ameliorate the course of the illness and prevent amyloidosis.

Clinical Management

Goldfinger (1972), Wolff et al. (1974), and Ravid et al. (1977) reported benefit from colchicine in reducing painful attacks in FMF.

Zemer et al. (1986) presented evidence that colchicine prevents and ameliorates amyloidosis in FMF. They followed 1,070 patients with FMF for 4 to 11 years after they were advised to take colchicine to prevent febrile attacks. Overall, at the end of the study, the prevalence of nephropathy was one-third of that in a study conducted before colchicine was used to treat FMF. Among 960 patients who initially had no evidence of amyloidosis, proteinuria appeared in 4 who adhered to a prophylactic schedule and in 16 of 54 who admitted noncompliance. Life-table analysis showed that the cumulative rate of proteinuria was 1.7% after 11 years in the compliant patients and 48.9% after 9 years in the noncompliant patients. All 24 patients with nephrosis or uremia had progressive deterioration of renal function. In 86 patients with proteinuria but no nephrotic syndrome, proteinuria resolved in 5 and stabilized in 68 (for more than 8 years in 40 patients). In the experience of Knecht et al. (1985), patients in whom colchicine fails to prevent attacks and SAA spikes enjoy as effective protection against renal amyloidosis as do colchicine-responsive patients. Jones et al. (1977) and Benson et al. (1977) reported recurrence of amyloid in kidney after renal transplant in FMF.

Schwabe and Nishizawa (1987) described a 36-year-old male of pure Japanese ancestry with a classic 20-year history of recurrent FMF manifested by self-limited attacks of fever plus pleuritis, peritonitis, or arthritis. The attacks were completely suppressed by daily prophylactic colchicine but recurred when the drug was briefly discontinued. He had been free of attacks for 10 years while taking 1.2 mg of colchicine daily.

Zemer et al. (1993) reported a family in which 6 of 9 sibs had FMF, the oldest born in 1950 and the youngest in 1970. The youngest was brought for clinical examination at the age of 12 years by his 'painfully experienced and observant mother' because his urine looked suspicious, and proteinuria was found. With continuous colchicine treatment, proteinuria persisted for 3 years and then gradually subsided over the next 2 years. By the age of 22 years, his urine had been free of protein for 5 years.

Ben-Chetrit et al. (1996) measured simultaneous colchicine levels in serum and breast milk in 4 breast-feeding women with a diagnosis of FMF for over 7 years. The authors found that the levels of colchicine in serum and breast milk paralleled each other. The peak concentration of colchicine occurred between 1 to 3 hours in all women. They found no abnormalities in the 4 children after 10 months of follow-up. Although the authors postulated that colchicine did 'no harm to the breast feeding infant,' they also stated that some women may consider breastfeeding 12 hours after the colchicine has been ingested and bottle feed for the other 12 hours.

Milledge et al. (2002) serendipitously found that allogeneic bone marrow transplantation (BMT) can cure FMF. They described a 7-year-old girl with congenital dyserythropoietic anemia (CDA; see 224120) who also had FMF and carried the MEFV M680I mutation (608107.0004), which was inherited from her father. Repeated transfusions required since the age of 6 months to treat her CDA led to iron overload and a persistently high ferritin level. Relapsing FMF made effective iron chelation therapy very difficult. Consequently, at the age of 4 years, she underwent allogeneic BMT from her brother, who did not carry the M680I mutation. During conditioning for her BMT, symptoms of FMF, including splenomegaly, arthritis, and recurrent abdominal pain, began to resolve and she was gradually weaned off colchicine. Two years after the transplantation, she remained free from FMF symptoms and was off all immunosuppressants. Milledge et al. (2002) stated that the findings in this patient demonstrated that symptoms of FMF can be alleviated by the therapy used during allogeneic BMT, and that it was likely that the missing factor in FMF in this patient was provided by granulocytes derived

from the stem cells within transplanted bone marrow. Both the patient and her donor brother had 6 silent polymorphisms of the MEFV gene inherited from the mother. Touitou (2003) suggested that the 'assertion (that through BMT the missing factor in FMF was provided) should be tempered.' Touitou et al. (2003) maintained that BMT has no role in the treatment of FMF, given the morbidity and mortality associated with BMT, and given that FMF is a self-limited disease with an overall good prognosis when colchicine is properly adjusted. They suggested that even in the setting of a well-designed research study, the ethical issue of offering this high-risk, unproven procedure to children is particularly troubling.

Touitou et al. (2007) reported transmission of an FMF mutation via bone marrow transplantation between 2 brothers of Sephardic Jewish origin. One brother had idiopathic aplastic anemia and received allogenic bone marrow transplant from his identical brother at age 5 years. At age 6 years, the donor brother developed clinical features of FMF, which was confirmed by genetic analysis. The transplanted brother, who was born without the FMF mutation, was found to have circulating FMF mutated cells, which fluctuated from 77 to 86.5%. However, by age 8 years, the recipient had not developed symptoms of the disorder.

Pathogenesis

Babior and Matzner (1997) suggested that the pathogenesis of FMF is as follows: pyrin, or marenostrin, is postulated to activate the biosynthesis of a chemotactic factor inactivator, an enzyme that normally occurs in the serosal fluids. They suggested that a chemotactic factor (probably C5a; 113995) can be released by subclinical injury to the serosa during normal activities, but the amounts released are small enough that they are cleared by the inactivating enzyme before they can provoke an inflammatory reaction. In FMF the inactivating enzyme is absent, allowing the chemotactic factors to survive long enough to call in neutrophils, which then release a variety of products, including an enzyme that generates more C5a. The result is an upward spiral that culminates in a full-blown inflammatory reaction: an attack of FMF.

Mapping

In linkage studies in Armenians, Shohat et al. (1990) excluded FMF from those portions of the genome at least 15 cM from 14 genetic markers, and in other linkage studies, Shohat et al. (1990) concluded that the immunogenetic region of chromosome 6 could be excluded from linkage with FMF in Armenian families. By linkage analysis, Gruberg et al. (1991) excluded several candidate genes including lipocortins, dopamine beta-hydroxylase, and interleukins 1 and 6. Kastner et al. (1991) presented a 90-marker exclusion map. With marker D17S74 on chromosome 17, they obtained a maximum multipoint lod score of 3.54 approximately 15 cM telomeric to the marker.

Pras et al. (1992) succeeded in mapping the FMF gene to 16p by linkage studies in 27 non-Ashkenazi Jewish families in Israel. One DNA marker, D16S84, gave a maximum lod score of 9.17 at a recombination frequency of 0.04. A probe associated with the hemoglobin alpha complex (5-prime-HVR) gave a maximum lod score of 14.47 at theta = 0.06. Multipoint analysis indicated that the likely order is as follows: cen--FMF--D16S84--HBA--tel. The maximum multipoint lod score was 19.86. There was a striking degree of homozygosity at chromosome 16p loci in the affected offspring of 8 consanguineous couples, thus supporting linkage by the method of homozygosity mapping. Pras et al. (1992) and Fischel-Ghodsian et al. (1992) found that the FMF gene maps to 16p in all ethnic groups, including Armenians and Ashkenazi and non-Ashkenazi Jews.

Studying 14 Armenian and 9 non-Ashkenazi Jewish families with FMF, Shohat et al. (1992) found linkage to the alpha-globin complex on 16p in both groups, with no evidence for genetic heterogeneity either between the groups or within the groups. Aksentijevich et al. (1992, 1993) observed different haplotypes associated with the disease in strong linkage disequilibrium in Moroccan and Iraqi Jewish families. This, together with the fact that Moroccans had a much more severe form of FMF, suggested that the 2 groups carry different allelic mutations. The mutation in Armenians may also be different from that in Moroccans, accounting for the milder phenotype.

Aksentijevich et al. (1993) attempted more precise mapping of MEFV by the homozygosity method for 8 of 9 markers. The rate of homozygosity among 26 affected inbred individuals was higher than that among their 20 unaffected sibs. Localizing MEFV more precisely on the basis of homozygosity rates alone would be difficult, they concluded, for 2 reasons: the high FMF carrier frequency increases the chance that inbred offspring have the disease without being homozygous by descent at the MEFV locus; and several of the markers in the MEFV region are relatively nonpolymorphic, with a high rate of homozygosity, regardless of their chromosomal location.

Aksentijevich et al. (1993) reexamined the linkage data to ascertain the possible reason for the earlier conclusion that the gene causing FMF is on 17q (Kastner et al., 1991). They found that the data with the chromosome 17 markers alone suggested locus heterogeneity. Nonetheless, the families were not separable into complementary subgroups showing linkage either to chromosome 16 or to chromosome 17. They examined the possibility that the positive lod score for chromosome 17 might reflect a secondary, modifying locus. However, by several measures of disease severity, families with positive lod scores for chromosome 17 loci had no worse disease than those with negative lod scores for these loci. They concluded that chromosome 17 does not carry a major FMF susceptibility gene for some families and does not encode a disease-modifying gene. Rather, it appeared that linkage to chromosome 17 was a 'false positive' (type I) error. These results reemphasize the fact that a lod score of 3.0 corresponds to a posterior probability of linkage of 95%, with an attendant 1 in 20 chance of observing a false positive.

FMF in the Arab population is said to be characterized by a low incidence of arthritis, amyloidosis, and erysipeloid erythema. These differences in disease expression raise the possibility of locus heterogeneity. However, Pras et al. (1994) found that the FMF gene maps to 16p in Druze and Moslem Arab families. Shohat et al. (1992) had shown that the gene maps to the same region in Armenians and non-Ashkenazi Jews.

Using linkage disequilibrium mapping in the study of 65 Jewish, Armenian, and Arab families, Levy et al. (1996) obtained a maximum lod score of 49.2 at a location 1.6 cM centromeric to D16S246. A specific haplotype using 3 markers was found in 76% of Moroccan and 32% of non-Moroccan Jewish carrier chromosomes, but this haplotype was not overrepresented in Armenian or Arab FMF carriers. Since the Moroccan Jewish community represents a relatively recently established and genetically isolated founder population, Levy et al. (1996) analyzed the Moroccan linkage-disequilibrium data and placed the FMF susceptibility gene within 0.305 cM of D16S246.

The French FMF Consortium (1996) narrowed the location of the FMF gene to a 250-kb interval through the study of non-Ashkenazi Jewish founder haplotypes accomplished by use of 15 microsatellite markers. They concluded that the FMF locus is situated between D16S3070 and D16S3275. Sood et al. (1997) used a high-resolution clone map of the 16p13.3 region to narrow the FMF interval. They identified several founder haplotypes in various ethnic groups.

Akarsu et al. (1997) studied 8 consanguineous Turkish families with at least 2 offspring affected with FMF. In 6 of these families, linkage was observed with a maximum lod score of 9.115 at theta = 0.00 for marker D16S3024. Two families, however, were unlinked to this region. Haplotype construction showed homozygosity for the region bounded by D16S3070 and D16S2617 in 5 families; 80% allelic association was shown with D16S2617.

Molecular Genetics

By screening 165 individuals from 65 families with familial Mediterranean fever, the International FMF Consortium (1997) identified 3 different missense mutations in exon 10 of the MEFV gene (M694V; 608107.0001, V726A; 608107.0003, and M680I; 608107.0004) that accounted for 78 carrier chromosomes.

The French FMF Consortium (1997) identified 4 sequence variations (608107.0001-608107.0004) in the marenostrin gene that correlated with FMF in various ethnic groups. In 72% of the patients in their sample, 1 or 2 of the 4 mutations were found.

For a detailed discussion of GENOTYPE/PHENOTYPE CORRELATIONS, see 608107.

Cazeneuve et al. (2003) analyzed a group of 50 patients with FMF from Karabakh, where the population is mainly of Armenian descent, and reviewed published series of classically affected populations. They found that whereas the distribution of genotypes at the MEFV locus differed dramatically from the Hardy-Weinberg equilibrium (p = 0.0016 and p less than 0.00001, respectively), the distribution of the most common MEFV mutations did not. Based on these results and other population genetics-based data, Cazeneuve et al. (2003) suggested the existence of a novel FMF-like condition that, depending upon the patients' ancestry, would affect 85 to 99% of those with no identified mutation in the MEFV gene.

Population Genetics

For a detailed discussion of particular allele and mutation frequencies of the MEFV gene in different populations, see 608107.

Sohar et al. (1967) estimated that in some Jewish groups the frequency of familial Mediterranean fever is 1 in 2,720 and that the minimal estimates for gene frequency and heterozygote frequency are 1 in 52 and 1 in 26, respectively. The number of Ashkenazi cases observed in Israel by Sohar et al. (1967) explains that a fair number of cases are observed in the large Ashkenazi group in the United States. Familial Mediterranean fever occurs mainly in Armenians and Sephardic Jews (those who left Spain during the Inquisition and settled in various countries bordering the Mediterranean). The possibility that the disorder in Armenians is distinct from that in Sephardic Jews is suggested by the lower frequency of amyloidosis (Schwabe and Peters, 1974) and longer average survival. Schwabe et al. (1977) reported 197 patients: 131 Armenians, 11 Ashkenazim, 27 non-Ashkenazi Jews, and 28 others. In an analysis of 1,327 cases from the literature, Meyerhoff (1980) found that 50% were Sephardic, 22% Armenian, 11% Arabian, 7% Turkish, and 5% Ashkenazi. Rawashdeh and Majeed (1996) reviewed the incidence of amyloidosis complicating FMF among several Mediterranean ethnic groups.

Using extended pedigree data of 90 FMF probands, Daniels et al. (1995) calculated the FMF gene frequency in various ethnic groups in Israel by analyzing the frequency in a total of 2,312 first cousins. The heterozygote frequencies were as follows: 1 in 4.9 (0.2 +/- 0.06) for the Libyan subgroup; 1 in 6.4 (0.16 +/- 0.03) for the subgroup from other North African countries; 1 in 13.3 (0.07 +/- 0.04) for the Iraqi subgroup; 1 in 11.4 (0.09 +/- 0.06) for the Ashkenazi subgroup; and 1 in 29.4 (0.03 +/- 0.03) for the remaining ethnic groups. The observed number of affected parents and affected offspring of probands was in agreement with the estimated gene frequency. Pras et al. (1998) commented that North African Jews and Iraqi Jews were the 2 largest population groups suffering from FMF in Israel. North African Jews were found to have more severe disease manifested by earlier age of onset, increase in frequency and severity of joint involvement, higher incidence of erysipelas-like erythema, and higher dose of colchicine required to control symptoms.

El-Shanti et al. (2006) provided a detailed review of FMF in the Arab population.

In a metaanalysis based on literature reports of 16,756 chromosomes from FMF patients and normal individuals from 14 affected populations, Papadopoulos et al. (2008) demonstrated that MEFV mutations were not distributed uniformly along the Mediterranean Sea area. The most frequent mutations were M694V (39.6%), V726A (13.9%), M680I (11.4%), E148Q (608107.0005) (3.4%), and M694I (608107.0002) (2.9%). However, 28.8% of chromosomes did not have identified mutations, particularly among western Europeans. The mean overall carrier rate was 0.186 with peak values in Arab, Armenian, Jewish, and Turkish populations. The V726A mutation was the only mutation to show Hardy-Weinberg equilibrium, implying that this mutation is the most ancient. Individuals in the Jewish population presented the most intense genetic isolation and drift, suggesting that they might have nested de novo mutations and accelerated evolution. Additional population groups appeared to follow distinct evolutionary lines in Asia Minor, eastern Europe, and western Europe.

Bonyadi et al. (2009) studied 524 unrelated Iranian patients of Azeri Turkish origin with FMF and identified mutations in the MEFV gene in 57% of alleles. The R761H was the most frequent (4.7%) of the rare alleles, and the authors suggested that R761H should be included in routine molecular diagnosis of FMF patients from this ethnic group. Bonyadi et al. (2009) concluded that FMF is no longer a rare disease in Iran.

Among 15,854 ethnically diverse individuals screened for familial Mediterranean fever carrier status, Lazarin et al. (2013) identified 247 carriers (1.6%), for an estimated carrier frequency of 1 in 64. Three individuals were identified as homozygotes or compound heterozygotes. Five 'carrier couples' were identified. Of 745 individuals of southern European origin, a carrier frequency of 1 in 75 was found. Of 392 individuals of Middle Eastern origin, a carrier frequency of 1 in 25 was found.

History

Sack (1988) found novel structural changes in members of the serum amyloid A gene family in 4 FMF patients of varied ethnic backgrounds. He interpreted these observations as suggesting 'that alterations of serum amyloid A genes, their protein products, and/or their regulation may be responsible' for FMF. Shohat et al. (1989, 1990) and Sack et al. (1991) excluded close linkage between FMF and the SAA locus (104750), however. Similarly, Shohat et al. (1989, 1990) demonstrated that the gene for serum amyloid P component (104770) is not closely linked to the locus for FMF. The complication of amyloidosis varies in frequency among various ethnic groups. Although FMF might not be 'caused' primarily by a mutation in the SAA or SAP gene, proneness to amyloidosis might differ according to a particular polymorphic allele at one or the other locus carried by the ethnic group. The failure of Shohat et al. (1990) to find an association of a particular polymorphism with amyloidosis probably rules out this possibility. Shohat et al. (1989) advanced the hypothesis that FMF patients are homozygous for a mutant allele for one of the lipocortin genes (151690), resulting in either lipocortin deficiency or production of an abnormal lipocortin protein. Lipocortin may be especially critical during stress. Its deficiency could result in a lack of feedback inhibition and an increase in the release of arachidonic acid, precursor of the potent mediators of inflammation. The deficiency might result in increased generation of prostaglandins, leukotrienes, and other inflammatory mediators by granulocytes, which then further activate phospholipase A2 by a feedback mechanism.