Cardiovascular and Metabolic Risk Factors in Inherited Autoinflammation

Context: The natural progression of metabolic abnormalities among patients with inherited autoinflammation is unclear. Objective: The objective of the study was to assess the cardiometabolic risk of participants with familial Mediterranean fever (FMF). Design Setting: This study included nationwide cross-sectional and longitudinal cohorts. Participants: The prevalence of components of the metabolic syndrome at age 17 was as-sessed from the medical database of the Israeli Defense Force from 1973 through 1997. were 745 males of 22 tables. A t test was used to test mean differences in continuous variables between FMF and controls. Multinomial logistic regression analysis, with normal weight or BP as the base category for comparison, was used to assess the association between FMF and abnormal BMI and BP, respectively. Cox pro-portional hazard models were used to estimate the hazard ratio and 95% confidence intervals (CI) for the incidence of the metabolic outcomes at young adulthood using the control partici-pant group as a reference. All tests used were two tailed, and P (cid:3) .05 was considered statistically significant. Analyses were per-formed with IBM SPSS, version 19.

Conclusions: FMF is associated with lower rates of most components of the metabolic syndrome compared with normal subjects, unlike other inflammatory conditions. (J Clin Endocrinol Metab 99: E2123-E2128, 2014) F amilial Mediterranean fever (FMF) is the most frequent genetic autoinflammatory disease (1) and is characterized by recurrent episodes of fever and serositis (1). Mutations in MEditerrenean FeVer (MEFV) lead to acute episodes of inflammation caused by activation of an inflammasome, which result in spontaneous release of IL-1␤, a key inflammatory cytokine in the pathogenesis of FMF (2) that was associated with deleterious cardiometabolic outcomes (3,4). Current data on cardiometabolic risk in FMF are ambiguous (5)(6)(7). We report on the incidence of cardiometabolic risk factors among FMF patients using a large cohort of military personnel of the Israeli Defense Forces (IDF).

Study population
The study group consisted of all 17-year-old recruits to military service in the IDF from 1973 through 1997. All eligible Israeli adolescents undergo a thorough medical evaluation as detailed elsewhere (8). Subjects were assigned to the FMF group based on a documented diagnosis by a rheumatologist at the time of medical evaluation, whereas the non-FMF subjects were considered controls (8). A second, separate control group was composed of healthy male siblings of FMF patients. Prospective analysis of the incidence of cardiovascular and metabolic risk factors was based on the medical records of subjects in the above cohort who remained in service as career army personnel. All army personnel older than 25 years of age are referred every 5 years for a routine health examination and screening tests at the Staff Periodic Examination Center (SPEC) as part of the Metabolic, Lifestyle, and Nutrition Assessment in Young Adults (MELANY) cohort (9). At each visit, the participants complete a detailed questionnaire assessing demographic, nutritional, lifestyle, and medical factors and provide blood samples after a 14-hour fast as reported elsewhere (10).
Although a normal creatinine clearance is a prerequisite for recruitment to the army, the diagnosis of FMF does not disqualify a subject from a future army career. Indeed, the overall prevalence of individuals with FMF among career army personnel is even higher than in the general population (0.42% vs 0.11%, respectively).

Inclusion/exclusion criteria
To assess the odds ratio (OR) of FMF patients for abnormal body mass index (BMI) or blood pressure (BP) measurements at adolescence, we included participants aged 16 -20 years at the time of enrollment (medical examination at age 17 y) from 1973 through 1997, who were medically eligible for military service. Subjects who were ineligible for military service due to any preexisting medical conditions were excluded from analysis, as reported previously (8) (See Supplemental Figure 1 for details). The current analysis included 788 459 male adolescents (745 with a diagnosis of FMF). No data were available on the rate of colchicine treatment for the FMF group at the time of enrollment. A total of 902 healthy male siblings (related to 593 FMF participants) met the inclusion criteria.
From the above cohort, we prospectively followed up those who became career army personnel and underwent periodic medical screening (the MELANY cohort) with at least one SPEC visit from 1994 through 2011. Because an individual with a diagnosis of FMF is limited to an office-based army service, only office-based army personnel were included in the control group. Excluded from the analysis were subjects who at enrollment had a diagnosis of any dyslipidemia, diabetes, diagnosis of hypertension, and a follow-up period less than 10 years. Included in the final analysis were 1625 men (57 with a diagnosis of FMF and under colchicine treatment). Only four women with FMF at the MELANY cohort met the above inclusion criteria; therefore, the analysis was limited to men only. The Institutional Review Board of the IDF approved the study and waived the requirement for informed consent on the basis of strict maintenance of participants' anonymity.

Study outcomes and follow-up
Supplemental Figure 1 outlines a summary of the design and outcome of the study. Cross-sectional analysis included the occurrence of overweight, obesity, prehypertension (BP systolic Ͻ 120/Ͻ 140 mm Hg or BP diastolic Ͻ 80/Ͻ 90 mm Hg) and hypertension-range (BP systolic Ն 140 mm Hg or BP diastolic Ն 90 mm Hg) measurements.
The incidence of cardiovascular and metabolic outcomes as measured by routine SPEC visits included the following: overweight BMI Ն 25 kg/m 2 , obesity (BMI Ն 30 kg/m 2 ), impaired fasting glucose (glucose Ն 100 mg/dL), diabetes, elevated BP (BP systolic Ն 130 mm Hg or BP diastolic Ն 85 mm Hg or diagnosis of hypertension), increased low-density cholesterol (LDL) level (Ն130 mg/dL), hypertriglyceridemia (Ն150 mg/dL), or incident low high-density cholesterol (HDL) (Յ40 mg/dL). Follow-up time was measured separately for each outcome and ended at the occurrence of death, discharge from the army, age of 45 years, or March 1, 2011, whichever came first. None of the study participants were lost to follow-up.

Study variables
Socioeconomic status (SES) (low/medium/high), area of residence (urban/rural), education (Յ/Ͼ10 y), country of origin, physical activity (inactive/Ͻ150 min/wk/Ն150 min/wk), and smoking status (never/ex-smoker/current) were categorized as reported previously (8,11). BMI at adolescence was a categorical variable stratified according to the established percentile cutoff points by the Centers for Disease Control and Prevention for age (in months) and sex (12).

Statistical analysis
Categorical variables were compared between FMF and control group participants with 2 test and Fisher's exact test in case of 22 tables. A t test was used to test mean differences in continuous variables between FMF and controls. Multinomial logistic regression analysis, with normal weight or BP as the base category for comparison, was used to assess the association between FMF and abnormal BMI and BP, respectively. Cox proportional hazard models were used to estimate the hazard ratio and 95% confidence intervals (CI) for the incidence of the metabolic outcomes at young adulthood using the control participant group as a reference. All tests used were two tailed, and P Ͻ .05 was considered statistically significant. Analyses were performed with IBM SPSS, version 19.

Abnormal weight and BP at adolescence
The baseline characteristics of the study cohort are presented in Table 1. Participants with FMF had lower BMI values (Table 1), with 5.7% of these subjects overweight compared with 9.7% in the control group. In multinomial, multivariable logistic regression analysis, FMF had an OR of 0.65 for occurrence of overweight at adolescence (95% CI 0.44 -0.96, P ϭ .03; Figure 1A), whereas their healthy siblings tended to obesity (OR 1.48; 95% CI 1.04 -2.11, P ϭ .008; Figure 1A). Adolescents with FMF had lower BP compared with controls (Table 1), with an adjusted OR of 0.72 (0.61-0.85, P Ͻ .001) and 0.66 (0.48 -0.92, P ϭ .012) for prehypertension-and hypertension-range measurements, respectively ( Figure 1A). Table 1 also presents the characteristics of the 1625 participants at their first SPEC evaluation. Characteristics of each outcome are shown in Supplemental Table 1. Figure 1B shows univariate and multivariable analysis for each outcome. There were 1265 participants with normal BMI at enrollment (55 with FMF). A Cox regression multivariable model adjusted for age, birth year, BMI at adolescence, education, SES, country of origin, and physical activity yielded a hazard ratio of 0.32 (95% CI 0.10 -0.82, P ϭ .002) for incident obesity and 0.62 (95% CI 0.39 -0.98, P ϭ .048) for incident overweight at adulthood. Adults with FMF had lower systolic, but not diastolic, BP compared with controls (Table 1); there was a nearly 2-fold increase in incidence rate of elevated BP measurement, which became statistically insignificant in multivariable model ( Figure 1B).

Cardiometabolic risk factors during adulthood
The FMF group displayed a distinct plasma lipid profile that was characterized by lower HDL, LDL, and triglyceride levels compared with controls (Table 1). This differences persisted in multivariable analysis ( Figure 1B and Supplemental Table 1) and were characterized by an older age of onset among the FMF group compared with controls (P Ͻ .048; Supplemental Table 1).

Discussion
The cardiovascular risk associated with FMF is undetermined (5, 6). Here we found that FMF patients with normal renal status had lower rates of most components of the Figure 1. The incidence of the metabolic syndrome components at adolescence and adulthood among men with FMF. A, The occurrence of abnormal BMI and BP measurements at adolescence. Overweight and obese correspond to the 85th-95th and greater than the 95th percentile of age-adjusted Centers for Disease Control and Prevention tables. Multivariable models for overweight and obesity were adjusted for age, birth year, country of origin, education, SES, and area of residence with controls (no FMF or a sibling with FMF) as the reference group. BMI was added to the model for BP outcomes. Note that the 95% CI for FMF participants and their healthy siblings do not overlap for overweight, obesity, and prehypertension outcomes. Prehypertension range, systolic BP 120 to 140 or diastolic BP 80 to 90; hypertension range 140 or greater/90 mm Hg or greater. B, The incidence of metabolic syndrome components during adulthood for 1568 control and 57 FMF participants. Univariate (adjusted for birth year and age at enrollment) and multivariable models (adjusted to age, birth year, education, SES, country of origin, physical activity; BMI was added to the model when it was not an outcome) are shown with 95% CI. Note that for outcomes including incident overweight or obesity, only normal-weight participants were included. Follow-up data for each outcome are available in the supplements (Supplemental Table 1). TG, fasting plasma triglycerides.

E2126
Twig metabolic syndrome yet also a lower HDL with comparable rates of dysglycemia ( Figure 1B). Among the FMF group, a lower prevalence of obesity persisted from adolescence throughout the entire follow-up despite a higher prevalence (Ͼ80%) of physical inactivity (Table 1). Previous studies showed that increased BMI at adolescence, already within the high-normal range, is an independent risk factor for cardiometabolic comorbidities and mortality (9,12,13). Overweight at adolescence was also associated with an increased risk for end-stage renal disease independent of diabetes status and BP at baseline (14). Indeed, the observation that lower incidence rates of obesity and elevated BP persisted throughout adulthood is concordant with an undetected fraction of nonamyloidosis-related end-stage renal disease during 17 000 person-years of follow-up on men with FMF (8). Although paroxysmal abdominal pain may contribute to the differences in BMI, it is unlikely that chronic illness is the sole explanation for the lower BMI among FMF participants because more than 40% of them became overweight or obese at adulthood (Supplemental Table 1). The latter factor also suggests that men with FMF are only partially protected from the growing obesity epidemic of Western society, including Israel (15).
Ongoing inflammatory state similar to that observed in rheumatoid arthritis was associated with insulin resistance and with decreased total cholesterol and HDL levels (16). Such a mechanism provides a possible explanation for the comparable dysglycemia rates in the presence of BMI differences. However, the FMF group in our study was characterized by lower triglyceride levels (Table 1 and Figure 1B), as opposed to the reported hypertriglyceridemia in a variety of other inflammatory conditions (16). The latter difference persisted after adjustment for BMI and can be another component that is unique to the inflammatory process of FMF (17).
Colchicine is the drug of choice in FMF and can potentially affect some of the studied outcomes. Colchicine was shown recently to benefit patients with stable angina from recurrent cardiovascular events (18), even though it was reported to have no effect on plasma lipids levels (19). It was also suggested to impair insulin release (20), thereby providing another explanation why the lower BMI did not protect from incident dysglycemia. The adverse gastrointestinal effect of chronic colchicine use may lower total caloric intake. However, no obesity-protecting effect of chronic colchicine treatment among adults with gout or relapsing pericarditis was ever reported.
Several limitations of this study warrant consideration. First, the military participants with FMF may be considered healthier than individuals with FMF who are not in military service. To attenuate this potential bias, we in-cluded as controls army personnel who were enrolled in the same type of service, with careful consideration of potential confounders. The relatively homogeneous environment with free health care service to which career army personnel are exposed might be advantageous by minimizing the effect of socioeconomic-dependent and unrecognized confounders. Second, the absence of direct measurement of insulin sensitivity limited our ability to evaluate the nature of dysglycemia in the FMF group. Inflammatory markers and underlying genetics were not available, limiting our ability to stratify the findings by the degree of disease activity and genetic variants.
To conclude, the current cardiometabolic profile of young men with FMF demonstrated low rates of incident overweight and obesity, LDL, HDL, abnormal BP, and hypertriglyceridemia but comparable rates of dysglycemia. This profile differentiates FMF patients from those with other inflammatory conditions and supports the contribution of other mechanism(s) than inflammation per se. Future studies are needed to determine whether the different metabolic profile of patients with FMF is protective from cardiovascular events.