Relationship between clinical findings and genetic mutations in patients with familial Mediterranean fever

Familial Mediterranean Fever (FMF) is the most frequent autoinflammatory disease in Turkey [1, 2]. It is commonly seen among Turkish, Armenian, Arabic and Sephardic Jewish populations [3].

The most noticable finding of the disease is painful febrile episodes accompanied by peritonitis, pleuritic pain or acute synovitis in many patients. Headache or diffuse myalgia and malaise can also occur. The acute phase reactants like serum amyloid A (SAA) are significantly elevated [4, 5]. Recurrent orchitis and meningitis are uncommon presentations.

To date, many mutations on the responsible pyrin gene have been reported [6]. The pyrin gene was mapped on the short arm of chromosome 16 and contains approximately 60 kb sized genomic DNA [7].

In trials conducted in Turkish population, the most common mutations were reported as M694V, E148Q, M680I and V726A [8]. The last ten years have seen much research into the effect of particular genetic mutations on clinical findings [9].

In this study, we aimed to investigate the correlation between genetic mutations and clinical findings on admission of 562 patients with recurrent fever, abdominal pain and/or arthritis/arthralgia and subsequently diagnosed as having FMF. The relationship between genetic mutations, types and the severity of clinical findings in patients who were diagnosed as having FMF were analyzed. Furthermore, the relationship between frequency of proteinuria and genetic mutations were studied and the effect of colchicine on SAA was demonstrated.

Five hundred sixty-two children (52.8 %) were diagnosed as having FMF among the 1064 patients who presented with recurrent fever, abdominal pain, and arthritis/arthralgia and skin eruption (Fig. 1).

In this study we aimed to screen the correlation between genetic and phenotype of Turkish children who had been diagnosed as having FMF. Therefore, we investigated the 12 most frequent mutations and concomitant clinical and laboratory findings. The frequency of FMF among patients who were admitted with recurrent fever, abdominal pain, and arthritis/arthralgia and skin eruptions was 53.6 %. This result was similar to the study conducted by Ulgenalp et al. (50.5 %) [14].

The most common symptom was fever (97.3 %). The other frequent symptoms were abdominal pain (96.6 %) and arthritis/arthralgia (43.2 %/63.7 %). These clinical findings were mostly accompanied by M694V homozygous/heterozygous mutations. Similar to other studies, the most commonly seen mutation that accompanied clinical findings was M694V. Özelkaya et al. showed that the most frequent symptoms were abdominal pain (83.1 %) and fever (55 %). Other common findings reported are arthritis (17.1 %) and erysipelas-like erythema (7.7 %). It was also demonstrated that these clinical findings were mostly accompanied by M694V homozygote mutations [15‐17].

Solak et al. reported the most common symptom to be abdominal pain, which was followed by fever, arthralgia, chest pain, and skin eruptions in a study of 165 patients. In this study, the most commonly seen mutations were M694V/I, E148Q and M680I [18]. Akar et al. reported that the most frequent mutations were M694V and M680I [19]. Giancane et al. reported that the E148Q variant is common, of unknown pathogenic significance and as the only MEFV variant does not support the diagnosis of FMF [20].

In our study, distinct from other studies, E148Q was the second most common mutation that accompanied fever, arthritis, and arthralgia. M680I was the third most common mutation and was found less frequently in patients with arthralgia/arthritis, but was seen more frquently in patients with skin eruptions.

Chest pain was most frequently seen in patients with M694V homozygous and E148Q heterozygous mutations. Similar to other studies, the M694V was the most commonly determined mutation in these patients.

Marek et al. researched the second other mutation apart from 12 mutations in patients with heterozygous MEFV gene mutations and demonstrated the second mutation in 25 % of patients [21]. We did not investigate second analysis in patients with heterozygous mutations.

The mean age of patients who only had fever on admission was significantly lower than patients who presented with abdominal pain alone. It has been reported that fever is a common finding in younger patients; however, serositis, skin eruptions and other findings are more often seen in older patients with FMF [22]. Padeh et al. also stated that the first symptom of the disease might be fever and serositis may develop later in the course [23]. MEFV gene mutations can be found in patients with HSP. FMF has been reported more frequently in patients who had HSP than in the general population. MEFV mutations may be related to HSP susceptibility in children [24]. Salah et al. found heterozygous mutations of V726A and E148Q to be the most common in children who also had HSP [25]. In contrast, we found homozygous mutations of M694V and E148Q to be the most common. Dogan et al. found that 11 of their 76 patients (14.4 %) were heterozygous, and 5 patients (6.6 %) were homozygous and 2 were compound heterozygous [26]. Altug et al., found that 18 of their 68 patients with HSP had MEFV mutations, 15 of them being heterozygous [27].

Our study included more patients who were first diagnosed with FMF and later went on to have HSP. This could be the reason why our study had more homozygous mutations in contrast to the findings of the three studies mentioned above [27].

In our study, in patients who had been previously diagnosed as having FMF and presented with findings pertinent to HSP, we found that the most common mutation of the 105 patients was M694V (n = 55, 52.3 %); the second most common was E148Q (n = 28, 26.6 %). Homozygous M694V and E148Q mutations were particularly common (p = 0.03).

FMF strip assay is a fast screening test; the most common major 12 mutations with known clinical meaning can be investigated. Screening can investigate only 12 mutations, whereas sequencing gives a result of all changes of exons 2,3,5 and 10 on MEFV gene that can be reviewed. More detailed sequencing tehcniques can reveal that clinically diagnosed patients who had none of the 12 most common mutations may actually have some of the less common mutations or novel mutations.

Proteinuria was reported to be positive most commonly with the M694V mutation and less frequent with E148Q [28].

Serum Amyloid A levels on admission with homozygous M694V mutation were significantly higher than with other mutations such as heterozygous and compound heterozygous mutations. Similar to the study of Berkun et al., the SAA level was higher in patients with homozygous M694V mutation [29]. According to a study conducted by Duzova et al., SAA was higher with homozygous and compound heterozygous mutations [30]. There was no significant difference in SAA level between mutation types in children with complete response. The level of SAA was significantly lower in patients who were completely responsive to colchicine than in partially responsive patients [31]. Similar to our study, the authors emphasized that administration of colchicine might be effective in diminishing the SAA level [31]. In our study, we showed a decrease in level of SAA in patients who had a complete response to colchicine therapy. The dose of colchicine was not increased.

A retrospective case series review was conducted of all MEFV gene mutation testing completed at UCLA Clinical Molecular Diagnostic Laboratory by Ong et al. to correlate specific genotypes with adverse phenotypes of affected populations residing in the Western United States, between February 2002, and February 2012. This was followed by a clinical chart review of all subjects who either had a single or double mutation. All 12 common mutations in the MEFV gene were analyzed and the M694V variant was found to be associated with an adverse FMF clinical outcome in the Armenian-American population, which manifested as earlier onset of disease, increased severity of disease, and renal amyloidosis [32].

In our study, M694V was found to be the most common mutation and it heralds more proteinuria and a more severe clinical course.

In summary, the frequency of FMF was 52.8 %. Fever was the most frequent symptom. The most common mutation was M694V. The mean age of patients who were admitted with abdominal pain alone was significantly higher than patients who presented with fever alone. Abdominal pain was significantly more frequent in patients with M694V and E148Q heterozygous mutations than children with R202Q and M680I heterozygous mutations. Arthritis/arthralgia were significantly more common in patients with homozygous M694V and E148Q mutations. Chest pain was more frequent with homozygous M694V (61.4 %) and heterozygous E148Q (25.6 %) mutations (p = 0.04). In 25 of the 229 children (10.9 %) with chest pain, pericardial effusion was documented at echocardiography. Skin eruption was most commonly seen in homozygous E148Q mutations. This was followed by heterozygous M694V and heterozygous M680I mutations. Homozygous M694V (n = 34, 32.3 %) and E148Q (n = 23, 22 %) mutations were significantly more frequent in patients with HSP (p = 0.03). Arthritis (n = 17, 7.0 %) and arthralgia (n = 24, 6.7 %) were seen more frequently in patients with compound heterozygous M694V/E148Q mutations, whereas M694V/R202Q coexistence caused skin eruptions (n = 16, 11.4 %) more frequently. The most frequent occurence of febrile episodes were in patients with M694V homozygous mutations (number of attacks = 4.4 ± 1.6/month). Patients with a history of appendectomy (n = 26) carried homozygous M694V (n = 16) and E148Q (n = 10) mutations. Diarrhea developed after colchicine therapy in 0.22 % of patients. The ratio of the clinically-severe patients was 24.1 % (n = 70) in children with homozygous M694V mutations. This ratio was significantly higher than other homozygous mutations. Proteinuria was detected most frequently in patients with homozygous M694V mutations. The SAA level in homozygous M694V mutations was significantly higher than in other mutations.

Our study was comprehensive and has demonstrated the correlation between the genotype and the clinical characteristics of FMF in Turkish children. Patients were classified and analyzed according to clinical, laboratory, and genotype features. However, a major limitation of our study is its retrospective design and the lack of a demographic analysis due to insufficient information about the place of birth of the patients.