Advanced glycation end products and the CALLY index reflect inflammatory burden in familial Mediterranean fever

Abstract

Familial Mediterranean Fever (FMF) is a prototypical autoinflammatory disease characterized by recurrent inflammatory attacks and persistent subclinical inflammation. Advanced glycation end products (AGEs) accumulate under conditions of chronic inflammation and oxidative stress and are linked to metabolic and cardiovascular complications. The C-reactive protein–albumin–lymphocyte (CALLY) index has been proposed as a composite biomarker reflecting inflammatory and nutritional status. However, the relationship between tissue AGE accumulation and the CALLY index in FMF remains unclear. This study aimed to evaluate tissue AGE accumulation using skin autofluorescence (AGE-SAF) in patients with FMF and to investigate its associations with the CALLY index, inflammatory markers, metabolic parameters, and clinical characteristics. This cross-sectional study included 87 FMF patients and 52 healthy controls. AGE-SAF was measured non-invasively using an AGE Reader™ device. Clinical and laboratory data were collected, and the CALLY index was calculated from C-reactive protein (CRP), albumin, and lymphocyte count. Non-parametric tests and Spearman’s correlation analyses were applied. AGE-SAF levels were significantly higher in FMF patients than controls (2.06 ± 0.44 vs. 1.70 ± 0.20 AU; p < 0.001). AGE-SAF was elevated in FMF and correlated with inflammatory and metabolic markers, while showing an inverse association with the CALLY index.

Introduction

Familial Mediterranean Fever (FMF) is increasingly recognized as a chronic inflammatory condition characterized not only by recurrent attacks but also by persistent subclinical inflammation that contributes to long-term complications, including endothelial dysfunction, AA amyloidosis, and increased cardiovascular risk^1–5. Despite this, reliable biomarkers that capture the cumulative inflammatory and metabolic burden in FMF remain limited, and current markers largely reflect short-term inflammatory activity rather than long-term disease impact.

Chronic inflammation promotes oxidative stress and metabolic dysregulation, both of which drive the formation and accumulation of advanced glycation end products (AGEs). AGEs are generated through non-enzymatic glycation and oxidation of proteins and lipids and progressively accumulate in tissues under sustained inflammatory conditions^6,7. Beyond passive accumulation, AGEs interact with the receptor for advanced glycation end products (RAGE), leading to activation of intracellular signaling pathways, particularly nuclear factor kappa B (NF-κB), which amplifies proinflammatory cytokine production and oxidative stress^8–11. This AGE–RAGE–NF-κB axis establishes a self-perpetuating cycle of inflammation and oxidative damage, contributing to endothelial dysfunction and vascular injury^8–12. Therefore, tissue AGE accumulation may represent an integrated marker of chronic inflammatory and metabolic stress.

Skin autofluorescence (SAF) is a validated, non-invasive method for assessing tissue AGE accumulation and may reflect long-term exposure to oxidative and metabolic stress^13,14. Previous studies have demonstrated that elevated AGE-SAF levels are associated with endothelial dysfunction, arterial stiffness, and increased cardiovascular risk in a variety of metabolic and inflammatory conditions^15–18. However, despite the central role of chronic inflammation in FMF, data on tissue AGE accumulation assessed by SAF remain limited.

Composite inflammatory biomarkers integrating multiple physiological domains have recently gained attention as tools for evaluating systemic inflammatory burden. The C-reactive protein–albumin–lymphocyte index (CALLY) index has been proposed as a novel biomarker reflecting inflammatory activity, nutritional status, and immune competence^19–22. Initially developed in oncological settings, the CALLY index has been shown to have prognostic value in various malignancies, where lower values are associated with poorer survival outcomes and higher disease burden^19,22. More recently, similar composite indices have been investigated in chronic inflammatory diseases such as rheumatoid arthritis and ankylosing spondylitis, where they have been associated with disease activity, systemic inflammation, and functional status^23–28. These findings suggest that composite indices may better capture the multidimensional nature of chronic inflammatory diseases compared with single biomarkers.

Nevertheless, the relationship between tissue AGE accumulation and the CALLY index has not yet been investigated in patients with FMF. Understanding the interaction between a long-term tissue-based marker of cumulative damage (AGE-SAF) and a dynamic composite inflammatory index (CALLY) may provide a more comprehensive assessment of disease burden.

Therefore, this study aimed to evaluate tissue AGE accumulation using SAF in patients with FMF and to investigate its association with the CALLY index, inflammatory markers, metabolic parameters, and clinical characteristics. To our knowledge, this is the first study to examine this relationship, providing novel insight into the inflammatory–metabolic burden in FMF.

Materials and methods

This single-center cross-sectional study was conducted at the Rheumatology Outpatient Clinic of Bursa City Hospital. The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Bursa City Hospital (Approval No: 2026-3/8; Date: 4 February 2026). The study was registered at ClinicalTrials.gov (Identifier: NCT07439341). All participants provided written informed consent before enrollment. A total of 87 patients with FMF, diagnosed according to the Tel Hashomer criteria, and 52 healthy controls were enrolled in the study. Patients with FMF were recruited during routine follow-up visits at the rheumatology outpatient clinic, while the control group consisted of healthy volunteers without known chronic inflammatory, autoimmune, metabolic, renal, malignant diseases, or regular medication use. Exclusion criteria included: patients receiving glucocorticoid therapy; those with diabetes mellitus, chronic kidney disease (stage ≥ 3), active infection, malignancy within the past 5 years, other autoimmune or autoinflammatory diseases; and dermatologic conditions affecting the measurement site. All FMF patients were assessed during attack-free periods.

Demographic and clinical characteristics of all FMF patients were recorded, including age, sex, body mass index (BMI), disease duration, number of attacks in the past three months, MEFV mutation status, colchicine use and dosage, biologic treatment (anakinra or canakinumab), presence of amyloidosis, and comorbid conditions. Disease activity was defined as the presence of at least one FMF attack within the previous three months. Colchicine use was categorized as either locally manufactured or imported formulations, due to potential differences in formulation characteristics and bioavailability.

Non-invasive assessment of tissue AGE accumulation was performed by SAF using the AGE Reader™ device (DiagnOptics Technologies B.V., Groningen, The Netherlands). Measurements were obtained from the volar surface of the right forearm under standardized conditions. Measurements were performed by trained personnel according to the manufacturer’s recommendations. Three consecutive measurements were obtained, and the mean value was used for statistical analysis. Results were reported in arbitrary units (AU) according to the manufacturer’s calibration system. SAF is measured by exposing the skin to ultraviolet light (at approximately 370 nm), which induces fluorescence emission that can be detected between 420 and 600 nm.

Laboratory parameters derived from routine outpatient evaluations included C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), serum albumin, lymphocyte count, neutrophil count, platelet count, hemoglobin, serum amyloid A (SAA), fasting glucose, glycated hemoglobin (HbA1c), lipid profile (total cholesterol, Low-density lipoprotein (LDL)-cholesterol, High-density lipoprotein(HDL)-cholesterol, triglycerides (TG)), ferritin, vitamin B12, vitamin D, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and serum creatinine. Complete blood count parameters, including lymphocyte count, neutrophil count, platelet count, and hemoglobin, were measured using an automated hematology analyzer (Sysmex XN-9100, Sysmex Corporation, Kobe, Japan). CRP concentrations were determined by an immunoturbidimetric method using a Roche Cobas c703 analyzer (Roche Diagnostics, Mannheim, Germany). Serum albumin was measured using the bromocresol green method. Total cholesterol and triglyceride levels were determined using enzymatic colorimetric assays, HDL cholesterol was measured by a direct polymer–polyanion method, and LDL cholesterol was measured using a direct assay on the Roche Cobas c703 platform. Serum creatinine was measured using the alkaline picrate method standardized to isotope dilution mass spectrometry (IDMS). ALT and AST activities were determined using IFCC-calibrated UV assays. HbA1c was measured by high-performance liquid chromatography (HPLC) using an Adams HA-8180 V analyzer (Arkray Menarini, Kyoto, Japan). Ferritin, vitamin B12, and 25-hydroxyvitamin D concentrations were measured using electrochemiluminescence immunoassays on a Roche Cobas e801 analyzer (Roche Diagnostics, Mannheim, Germany). SAA concentrations were measured by nephelometry using a BN nephelometric analyzer (Siemens Healthineers, Erlangen, Germany). All laboratory analyses were performed according to the manufacturers’ instructions and routine internal quality-control procedures of the central laboratory. The triglyceride-to-HDL cholesterol ratio (TG/HDL) was also calculated as a marker of metabolic risk.

The CALLY index was calculated using CRP, serum albumin, and lymphocyte count values obtained from routine laboratory evaluations performed during outpatient follow-up visits. The CALLY index was calculated according to the formula: CALLY = [Albumin (g/L) × Lymphocyte count (10³/µL)]/CRP (mg/L) The CALLY index is a composite biomarker integrating markers of systemic inflammation and nutritional status. Higher CRP levels reflect inflammatory activity, whereas serum albumin and lymphocyte counts are indicators of nutritional and immune status. Therefore, the CALLY index has been proposed as a composite biomarker integrating inflammatory, nutritional, and immune parameters, originally developed in oncological settings and more recently explored in chronic inflammatory diseases^{20–22,26–28}.

Additional composite indices were calculated to further assess the systemic inflammatory and metabolic burden. The systemic immune–inflammation index (SII) was calculated as platelet count × neutrophil count/lymphocyte count. The HALP index was calculated using the following formula: (hemoglobin × albumin × lymphocyte count)/platelet count.

To provide an overall estimate of systemic inflammatory activity, an exploratory inflammatory burden score (IBS) was developed based on three inflammatory parameters: CRP, SII, and the CALLY index. These markers were selected because they reflect different components of systemic inflammation. CRP represents acute-phase inflammatory activity, SII integrates neutrophil, platelet, and lymphocyte counts and reflects immune–inflammatory balance, whereas the CALLY index incorporates inflammatory and nutritional status through CRP, albumin, and lymphocyte levels.

In the FMF cohort, median values for each parameter were calculated and used as cohort-specific thresholds. One point was assigned for CRP values above the cohort median, SII values above the cohort median, and CALLY index values below the cohort median. The total IBS was calculated as the sum of these points, resulting in a score ranging from 0 to 3, with higher scores indicating greater systemic inflammatory burden.

A median-based approach was used to avoid reliance on predefined clinical cut-off values and to allow balanced categorization of variables within the study population. However, this approach may limit comparability across studies and reduce clinical interpretability. This composite score was developed for exploratory purposes and has not been previously validated; therefore, it should be considered a hypothesis-generating measure rather than a clinically established index.

To evaluate cumulative metabolic stress associated with cardiometabolic risk factors, an exploratory metabolic burden score was constructed using body mass index (BMI), TG/HDL, and HbA1c. These parameters were selected because they represent key domains of metabolic health, including obesity, lipid-related metabolic risk, and glycemic status.

Each of the following conditions was assigned one point: BMI ≥ 30 kg/m² (indicating obesity), TG/HDL ratio above the cohort median (reflecting adverse lipid metabolism), and HbA1c ≥ 5.7% (indicating impaired glucose metabolism or prediabetes). The total metabolic burden score therefore ranged from 0 to 3, with higher scores indicating greater metabolic burden. Similar to the inflammatory burden score, this index was constructed for exploratory analyses in the present study and has not been externally validated. Accordingly, this score should be interpreted as exploratory and hypothesis-generating.

Given the exploratory nature of the study, a formal a priori sample size calculation was not performed. The sample size was determined by the number of eligible patients available during the study period.

Statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA). Continuous variables were expressed as mean ± standard deviation or median (interquartile range), depending on the distribution of the data, while categorical variables were presented as counts and percentages.

The normality of continuous variables was assessed using the Kolmogorov–Smirnov test and visual inspection of histograms. As the assumption of normality was not met for most variables, non-parametric tests were applied. Comparisons between two independent groups were performed using the Mann–Whitney U test, whereas comparisons among more than two groups were conducted using the Kruskal–Wallis test. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate.

Correlations between continuous variables were evaluated using Spearman’s rank correlation analysis. To identify independent factors associated with AGE-SAF levels, multivariable linear regression analysis was performed. Variables showing a univariable association with AGE-SAF (p < 0.05) or considered clinically relevant based on prior literature were considered for inclusion in the regression model. To minimize the risk of overfitting, the number of variables included in the final model was limited relative to the sample size. The number of predictors relative to the sample size was carefully considered to improve model stability.

Potential multicollinearity among predictor variables was assessed using variance inflation factor (VIF) values. Because the CALLY index showed a skewed distribution, logarithmic transformation was applied before inclusion in the regression model. AGE-SAF values were retained in their original scale because residual diagnostics supported the use of linear regression without transformation.

The assumptions of linear regression, including linearity, homoscedasticity, and normality of residuals, were evaluated using residual plots and diagnostic statistics. Model assumptions were assessed, and no major violations were observed. Because multiple subgroup analyses were performed for exploratory purposes, no formal correction for multiple comparisons (e.g., Bonferroni or false discovery rate) was applied, and the results should therefore be interpreted cautiously, particularly for variables with small subgroup sample sizes. Given the exploratory nature of these analyses, the findings should therefore be interpreted with caution. A two-sided p-value < 0.05 was considered statistically significant.

Results

The demographic characteristics of the FMF and control groups are presented in Table 1. A significant difference was observed between groups in terms of age (p = 0.013). The mean age was higher in the control group than in the FMF group (41.23 ± 9.73 vs. 36.31 ± 12.86 years, respectively) (Table 1). Because age differed between groups, age was included as a covariate in the regression analyses.

Table 1.

Demographic parameters.

	Control group (n = 52)	FMF group (n = 87)	p
Age (years)
Mean ± SD	41.23 ± 9.73	36.31 ± 12.86	0.013
Median (Min–Max)	41.5 (21–65)	35 (19–62)
Gender, n (%)
Female	26 (50.0)	45 (51.7)	0.844
Male	26 (50.0)	42 (48.3)
BMI (kg/m2)
Mean ± SD	25.70 ± 3.12	27.59 ± 8.67	0.303
Median (Min–Max)	25 (19.6–33.2)	26.1 (15.1–83.4)

The sex distribution was similar between the two groups. In the control group, the proportion of women was 50.0% (n = 26) and the proportion of men was 50.0% (n = 26), whereas in the FMF group, the proportion of women was 51.7% (n = 45) and the proportion of men was 48.3% (n = 42). No significant difference was observed between the groups in terms of sex distribution (p = 0.844) (Table 1).

No significant difference in BMI was observed between the groups (p = 0.303). The mean BMI was 25.70 ± 3.12 kg/m² in the control group and 27.59 ± 8.67 kg/m² in the FMF group. Although BMI values were numerically higher in the FMF group, this difference did not reach statistical significance (Table 1).

In the FMF group, the mean disease duration was 4.55 ± 4.81 years and the median value was 3 years (1–20). The mean number of attacks in the last three months was 1.61 ± 1.60, with a median of 1 (0–6). Active disease (≥ 1 attack/3 months) was observed in 60.9% of patients (n = 53) (Table 2).

Table 2.

Clinical parameters.

FMF group, n = 87
Disease duration (years) Mean ± SD; Median (Min–Max)	4.55 ± 4.81; 3 (1–20)
Number of attacks in the last 3 months Mean ± SD; Median (Min–Max)	1.61 ± 1.60; 1 (0–6)
Disease and treatment characteristics n (%)	Active disease (≥ 1 attack/3 months):	53 (60.9)
	Use of biologic treatment	5 (5.7)
	Duration of biologic treatment (months) Mean ± SD; Median (Min–Max)	0.63 ± 4.10; 0 (0–36)
	Use of anakinra	3 (3.4)
	Use of canakinumab	2 (2.3)
	Amyloidosis	5 (5.7)
	Use of locally manufactured colchicine	63 (72.4)
	Locally manufactured colchicine dose (mg) Mean ± SD; Median (Min–Max)	0.80 ± 0.54; 1 (0–2)
	Use of imported colchicine	24 (27.6)
	Imported colchicine dose (mg) Mean ± SD; Median (Min–Max)	0.51 ± 0.86; 0 (0–2)
	Ineffectiveness of locally manufactured colchicine	25 (28.7)
	Anti-TNF treatment	1 (1.1)
	Colchicine side effects	15 (17.2)
	Ankylosing spondylitis	4 (4.6)
Genetic mutation characteristics n (%)	M694V heterozygous	52 (59.8)
	R202Q heterozygous	32 (36.8)
	V726A heterozygous	10 (11.5)
	E148Q heterozygous	13 (14.9)
	m680I homozygous	3 (3.4)
	K361S heterozygous	1 (1.1)
	M694V homozygous	9 (10.3)
	R202Q homozygous	7 (8.0)
	m680I heterozygous	12 (13.8)
	V726A homozygous	1 (1.1)
	Compound M694V/V726A	5 (5.7)
	Compound M694V/M680I	5 (5.7)
	Compound M680I/E148Q	1 (1.1)
	Compound M694V/E148Q	8 (9.2)
	M694V homozygous + R202Q homozygous	4 (4.6)
	Compound M680I/V726A	2 (2.3)
	Compound M694V/R202Q	30 (34.5)
	Compound M694V/R761H heterozygous	4 (4.6)
	R761H homozygous	1 (1.1)
	R761H heterozygous	5 (5.7)
	Other mutations	8 (9.2)
Presence of comorbidity n (%)	27 (31.0)
	Hypertension	15 (17.2)
	Coronary artery disease	6 (6.9)
	Atherosclerosis	4 (4.6)
	Other comorbidity	18 (20.7)
Number of comorbidities n (%)	0–1	76 (87.4)
	2 or more	11 (12.6)

When treatment characteristics were evaluated, the majority of patients were using locally manufactured colchicine (72.4%, n = 63), whereas a smaller proportion were using imported colchicine (27.6%, n = 24). The mean locally manufactured colchicine dose was 0.80 ± 0.54 mg, median 1 mg (0–2); the mean imported colchicine dose was 0.51 ± 0.86 mg, median 0 mg (0–2). Non-response to locally manufactured colchicine was observed in 28.7% of patients (n = 25). Colchicine-related adverse effects occurred in 17.2% (n = 15), while ankylosing spondylitis was present in 4.6% (n = 4) of patients (Table 2).

The use of biologic treatment was limited, and 5.7% of patients (n = 5) were receiving biologic treatment. The mean duration of biologic treatment was 0.63 ± 4.10 months, with a median of 0 months (0–36). Among the biological agents used, anakinra accounted for 3.4% (n = 3) and canakinumab for 2.3% (n = 2). The rate of patients receiving anti-TNF treatment was 1.1% (n = 1). In addition, amyloidosis, one of the complications related to FMF, was observed in 5.7% (n = 5) (Table 2).

When the distribution of genetic mutations was examined, the most common mutation was the M694V heterozygous variant (59.8%, n = 52). This was followed by R202Q heterozygous (36.8%, n = 32) and compound M694V/R202Q (34.5%, n = 30). M694V homozygous mutation was detected in 10.3% (n = 9) and R202Q homozygous mutation in 8.0% (n = 7). V726A heterozygous mutation was found in 11.5% (n = 10) and E148Q heterozygous mutation in 14.9% (n = 13). Among the rarer mutations were m680I heterozygous (13.8%, n = 12), compound M694V/E148Q (9.2%, n = 8), and other mutations (9.2%, n = 8). K361S heterozygous, V726A homozygous, and R761H homozygous mutations were detected at a rate of 1.1% (Table 2). In the assessment of comorbidities, at least one comorbid disease was found in 31.0% of patients (n = 27). The most common accompanying conditions were other non-cardiovascular chronic comorbidities (20.7%, n = 18) and hypertension (17.2%, n = 15). In addition, coronary artery disease was detected in 6.9% (n = 6) and atherosclerosis in 4.6% (n = 4). When the number of comorbidities was examined, most patients had 0–1 comorbidity (87.4%, n = 76), while two or more comorbidities were observed in 12.6% (n = 11) (Table 2).

When the relationship between AGE-SAF and the metabolic burden score was evaluated on a group basis, a weak but statistically significant positive correlation was found between the two variables in the FMF group (r = 0.217; p = 0.044). Accordingly, higher AGE-SAF levels were associated with higher metabolic burden scores in the FMF group (Tables 3 and 4).

Table 3.

Primary measurement parameter: AGE level.

	FMF group (n = 87)	Control group (n = 52)	p
	Mean ± SD; Median (Min–Max)	Mean ± SD; Median (Min–Max)
AGE-SAF	2.06 ± 0.44; 2 (1.3–3.6)	1.70 ± 0.20; 1.70 (1.2–2.1)	< 0.001
Metabolic burden score	1.10 ± 0.30; 1 (1–2)	0.95 ± 0.88; 1 (0–3)	0.131

Table 4.

AGE-SAF and metabolic burden score.

		AGE-SAF
		r	p
Metabolic burden score	FMF group	0.217	0.044
	Control group	− 0.053	0.709

In contrast, AGE-SAF levels were not significantly correlated with the metabolic burden score in the control group (r = − 0.053, p = 0.709). Although the negative correlation coefficient suggested an inverse relationship, this association was weak and not statistically significant (Tables 3 and 4). These findings indicate that the association between AGE-SAF and metabolic burden was observed in the FMF group but not in the control group (Tables 3 and 4).

Statistically significant relationships were found between AGE-SAF and some metabolic and inflammatory parameters. AGE-SAF levels showed positive correlations with CRP (r = 0.335, p = 0.002), HbA1c (r = 0.285, p = 0.007), total cholesterol (r = 0.229, p = 0.033), triglycerides (r = 0.291, p = 0.006), and the TG/HDL ratio (r = 0.222, p = 0.039). In addition, significant positive associations were also observed between AGE-SAF and the number of attacks in the last three months (r = 0.315; p = 0.003) and the number of comorbidities (r = 0.227; p = 0.035). In contrast, a significant negative correlation was observed between AGE-SAF and the CALLY index (r=−0.321; p = 0.002). No significant associations were observed between AGE-SAF and other biochemical and clinical parameters (p > 0.05) (Tables 5 and 6).

Table 5.

Biochemical and laboratory parameters.

	Mean ± SD; Median (Min–Max)
CRP (mg/L)	10.82 ± 27.06; 2.57 (0.50–208)
Albumin (g/L)	46.00 ± 2.89; 46 (36.10–54.60)
Lymphocyte (103/µL)	2.23 ± 0.64; 2.14 (0.39–3.95)
Neutrophil (103/µL)	4.59 ± 1.67; 4.24 (1.92–11.91)
Platelet (103/µL)	263.76 ± 81.44; 264 (16.80–480)
Hemoglobin (g/dL)	13.9 ± 1.6; 13.9 (8–19.2)
ESR (mm/h)	10.69 ± 11.03; 8 (2–60)
SAA	8.92 ± 31.42; 1 (0–194)
HbA1c (%)	5.42 ± 0.68; 5.20 (4.40–8.80)
Total cholesterol (mg/dL)	176.70 ± 40.27; 173 (85.70–277)
LDL (mg/dL)	101.95 ± 34.74; 92 (34–196)
HDL (mg/dL)	45.88 ± 10.61; 44 (24.90–81)
Triglycerides (mg/dL)	154.77 ± 128.70; 125 (42–1110)
TG/HDL ratio	3.56 ± 2.64; 2.89 (0.93–15.42)
Vitamin D (ng/mL)	17.24 ± 10.47; 14.40 (2–51)
Vitamin B12 (pg/mL)	373.40 ± 238.90; 341 (153–2228)
Spot protein	241.06 ± 483.20; 114 (25–3404)
Spot creatinine	125.00 ± 73.88; 120 (13–355)
UPCR	1.89 ± 3.41; 1.10 (0.14–28.37)
Serum creatinine (mg/dL)	0.77 ± 0.25; 0.72 (0.49–2.14)
Ferritin (ng/mL)	84.80 ± 79.67; 63 (8.04–356)
ALT (U/L)	33.32 ± 24.20; 25 (6–126)
AST (U/L)	27.41 ± 22.53; 21 (10–179)
CALLY	65.20 ± 70.96; 43.29 (0.41–262.45)
HALP	25.93 ± 186.40; 5.29 (1.31–1744.08)
SII	593.87 ± 421.76; 512.20 (47.60–3558.90)
Inflammatory burden score	1.48 ± 1.25; 1 (0–3)

Table 6.

Relationships between AGE-SAF and CALLY indices and clinical/laboratory parameters.

	AGE-SAF		CALLY
	r	p	r	p
CALLY	− 0.321	0.002
CRP (mg/L)	0.335	0.002	− 0.971	< 0.001
Albumin (g/L)	− 0.131	0.226	0.173	0.110
Lymphocyte (103/µL)	− 0.079	0.469	0.270	0.011
Neutrophil (103/µL)	− 0.079	0.466	− 0.277	0.009
Platelet (103/µL)	− 0.055	0.612	− 0.169	0.117
Hemoglobin (g/dL)	− 0.003	0.979	− 0.012	0.914
ESR (mm/h)	0.184	0.088	− 0.573	< 0.001
SAA	0.199	0.065	− 0.744	< 0.001
HbA1c (%)	0.285	0.007	0.004	0.971
Total cholesterol (mg/dL)	0.229	0.033	0.036	0.738
LDL (mg/dL)	0.154	0.154	− 0.015	0.892
HDL (mg/dL)	0.001	0.996	0.126	0.244
Triglycerides (mg/dL)	0.291	0.006	− 0.011	0.923
TG/HDL ratio	0.222	0.039	− 0.060	0.579
Vitamin D (ng/mL)	0.085	0.432	− 0.029	0.789
Vitamin B12 (pg/mL)	0.191	0.077	− 0.210	0.051
Spot protein	0.033	0.764	− 0.042	0.702
Spot creatinine	− 0.004	0.969	− 0.041	0.709
UPCR	0.046	0.670	− 0.094	0.384
Serum creatinine (mg/dL)	0.100	0.356	− 0.113	0.299
Ferritin (ng/mL)	0.187	0.083	− 0.395	< 0.001
ALT (U/L)	0.121	0.264	0.087	0.423
AST (U/L)	0.162	0.135	0.111	0.308
HALP	− 0.016	0.883	0.379	< 0.001
SII	− 0.032	0.767	− 0.440	< 0.001
Inflammatory burden score	0.180	0.095	− 0.847	< 0.001
Locally manufactured colchicine dose (mg)	0.119	0.272	< 0.001	1.000
Imported colchicine dose (mg)	− 0.057	0.600	− 0.183	0.090
Duration of biologic treatment (months)	0.178	0.100	− 0.050	0.643
Disease duration (years)	− 0.079	0.470	0.139	0.200
Number of attacks in last 3 months	0.315	0.003	− 0.702	< 0.001
Number of comorbidities	0.227	0.035	− 0.095	0.382

Strong associations were observed between the CALLY index and several inflammatory markers. Significant negative correlations were found between CALLY and CRP (r=−0.971; p < 0.001), ESR (r=−0.573; p < 0.001), SAA (r=−0.744; p < 0.001), ferritin (r=−0.395; p < 0.001), SII (r=−0.440; p < 0.001), Inflammatory burden score (r=−0.847; p < 0.001), and the number of attacks in the last three months (r=−0.702; p < 0.001). In addition, a positive correlation was found between CALLY and lymphocyte count (r = 0.270; p = 0.011), and a negative correlation was found with neutrophil count (r=−0.277; p = 0.009). There was also a significant positive relationship between CALLY and the HALP index (r = 0.379; p < 0.001) (Table 6). No statistically significant relationship was found between the CALLY index and other laboratory and clinical variables (p > 0.05). These variables included albumin, platelets, hemoglobin, some lipid parameters (LDL, HDL), vitamin D, spot protein, spot creatinine, serum creatinine, ALT, AST, colchicine doses, duration of biologic treatment, disease duration, and number of comorbidities. In addition, there was a negative correlation between vitamin B12 and the CALLY index; however, this did not reach statistical significance (r = −0.210, p = 0.051) (Table 6). The strong inverse correlation between the CALLY index and CRP likely reflects the mathematical coupling between these variables, as CRP is a direct component of the CALLY formula. Therefore, this relationship should not be interpreted as an independent biological association. Overall, these findings indicate modest associations between AGE-SAF and several inflammatory, metabolic, and clinical parameters; however, they should be interpreted cautiously in view of multiple testing and the borderline significance of some correlations.

In the FMF group, the AGE-SAF levels differed according to several clinical characteristics. In patients who had experienced an attack within the last three months, the AGE-SAF levels were 2.18 ± 0.45; median 2.10 (1.40–3.60), whereas in patients who had not experienced an attack, it was 1.85 ± 0.35; median 1.85 (1.30–2.90), and this difference was statistically significant (p < 0.001). Similarly, in patients with active disease, the AGE-SAF levels were 2.18 ± 0.45; median 2.10 (1.40–3.60), whereas in those without active disease it was 1.89 ± 0.38; median 1.90 (1.30–2.90), and the difference was significant (p = 0.002) (Table 7).

Table 7.

Comparison of AGE-SAF levels according to clinical features.

		AGE-SAF		AGE-SAF	p
Attack in last 3 months	No (n = 32)	1.85 ± 0.35; 1.85 (1.30–2.90)	Yes (n = 55)	2.18 ± 0.45; 2.10 (1.40–3.60)	< 0.001
Active disease	No (n = 34)	1.89 ± 0.38; 1.90 (1.30–2.90)	≥ 1 attack/3 months (n = 53)	2.18 ± 0.45; 2.10 (1.40–3.60)	0.002
Biologic treatment	No (n = 82)	2.04 ± 0.43; 2 (1.30–3.60)	Yes (n = 5)	2.46 ± 0.47; 2.30 (1.90–3)	0.050
Use of anakinra	No (n = 84)	2.05 ± 0.44; 2 (1.30–3.60)	Yes (n = 3)*	2.50 ± 0.44; 2.30 (2.20–3)	0.055
Use of canakinumab	No (n = 85)	2.05 ± 0.44; 2 (1.30–3.60)	Yes (n = 2)*	2.40 ± 0.71; 2.40 (1.90–2.90)	0.486
Amyloidosis	No (n = 82)	2.04 ± 0.44; 2 (1.30–3.60)	Yes (n = 5)	2.44 ± 0.44; 2.30 (1.90–2.90)	0.054
Use of locally manufactured colchicine	No (n = 24)	1.96 ± 0.45; 1.90 (1.30–3.01)	Yes (n = 63)	2.10 ± 0.44; 2 (1.40–3.60)	0.155
Use of imported colchicine	No (n = 63)	2.09 ± 0.43; 2 (1.40–3.60)	Yes (n = 24)	2 ± 0.47; 1.95 (1.30–3.01)	0.360
Ineffectiveness of locally manufactured colchicine	No (n = 62)	2.06 ± 0.41; 2 (1.40–3.30)	Yes (n = 25)	2.07 ± 0.52; 2 (1.30–3.60)	0.944
Ankylosing spondylitis	No (n = 83)	2.07 ± 0.45; 2 (1.30–3.60)	Yes (n = 4)*	1.90 ± 0.36; 2 (1.40–2.20)	0.632
Other mutation	No (n = 79)	2.04 ± 0.39; 2 (1.30–3.01)	Yes (n = 8)	2.30 ± 0.78; 2.15 (1.40–3.60)	0.461
Anti-TNF treatment	No (n = 86)	2.07 ± 0.44; 2 (1.30–3.60)	Yes (n = 1)*	1.40	–

*Insufficient sample size.

AGE-SAF levels were higher in patients receiving biologic treatment than in those not receiving biologic treatment (2.46 ± 0.47; median 2.30 [1.90–3.00] vs. 2.04 ± 0.43; median 2.00 [1.30–3.60], respectively), although the difference did not reach statistical significance (p = 0.050). Similarly, although the AGE-SAF value appeared higher in patients using anakinra (2.50 ± 0.44; median 2.30), the difference did not reach statistical significance (p = 0.055). No significant difference in AGE-SAF levels was observed according to canakinumab use (p = 0.486) or the presence of amyloidosis (p = 0.054) (Table 7). When variables related to colchicine treatment were examined, no significant difference was found in AGE-SAF levels according to locally manufactured colchicine use (p = 0.155), imported colchicine use (p = 0.360), or ineffectiveness of locally manufactured colchicine (p = 0.944). In addition, AGE-SAF levels were similar according to the presence of ankylosing spondylitis (p = 0.632) and the presence of other mutations (p = 0.461). Since only one patient was receiving anti-TNF treatment, no statistical comparison was performed for this variable. These findings indicate that AGE-SAF levels are particularly associated with recent attacks and disease activity, whereas they did not show a significant relationship with most treatment and comorbidity variables (Table 7).

In the FMF group, the CALLY index was evaluated according to various clinical features. In patients who had experienced an attack within the last three months, the CALLY value was 32.73 ± 42.46; median 17.76 (0.41–169.77), whereas in patients who had not experienced an attack, it was 121.01 ± 75.77; median 96.86 (12.56–262.45), and this difference was statistically significant (p < 0.001). Similarly, in patients with active disease, the CALLY value was 32.33 ± 43.22; median 13.59 (0.41–169.77), whereas in patients without active disease it was 116.43 ± 75.76; median 88.24 (12.56–262.45), and the difference was statistically significant (p < 0.001) (Table 8).

Table 8.

Comparison of the CALLY index according to clinical features.

		CALLY		CALLY	p
Attack in last 3 months	No (n = 32)	121.01 ± 75.77; 96.86 (12.56–262.45)	Yes (n = 55)	32.73 ± 42.46; 17.76 (0.41–169.77)	< 0.001
Active disease	No (n = 34)	116.43 ± 75.76; 88.24 (12.56–262.45)	≥ 1 attack/3 months (n = 53)	32.33 ± 43.22; 13.59 (0.41–169.77)	< 0.001
Biologic treatment	No (n = 82)	67.41 ± 72.20; 43.37 (0.41–262.45)	Yes (n = 5)	29.01 ± 31.04; 18.66 (0.49–63.53)	0.232
Use of anakinra	No (n = 84)	66.74 ± 71.52; 43.37 (0.41–262.45)	Yes (n = 3)*	21.93 ± 36.03; 1.76 (0.49–63.53)	0.122
Use of canakinumab	No (n = 85)	65.80 ± 71.61; 43.29 (0.41–262.45)	Yes (n = 2)*	39.64 ± 29.67; 39.64 (18.66–60.62)	0.989
Amyloidosis	No (n = 82)	67.30 ± 72.29; 43.37 (0.41–262.45)	Yes (n = 5)	30.74 ± 29.15; 18.66 (3.93–63.53)	0.396
Use of locally manufactured colchicine	No (n = 24)	71.92 ± 86.23; 18.91 (0.49–262.45)	Yes (n = 63)	62.64 ± 64.82; 45.22 (0.41–241.86)	0.686
Use of imported colchicine	No (n = 63)	68.32 ± 68.76; 46.93 (0.41–262.45)	Yes (n = 24)	57.01 ± 77.36; 16.06 (0.49–238.74)	0.088
Ineffectiveness of locally manufactured colchicine	No (n = 62)	72.52 ± 70.02; 49.08 (0.41–262.45)	Yes (n = 25)	47.03 ± 71.39; 11.15 (0.49–238.74)	0.007
Ankylosing spondylitis	No (n = 83)	67.04 ± 72.05; 43.46 (0.41–262.45)	Yes (n = 4)*	26.97 ± 20.11; 27.93 (6.81–45.22)	0.378
Other mutation	No (n = 79)	60.19 ± 65.81; 42.77 (0.41–262.45)	Yes (n = 8)	114.63 ± 102.60; 120.46 (0.49–241.86)	0.329
Anti-TNF treatment	No (n = 86)	65.81 ± 71.14; 43.37 (0.41–262.45)	Yes (n = 1)*	12.56	–

*Insufficient sample size.

In patients using biologic treatment, the CALLY value was 29.01 ± 31.04; median 18.66 (0.49–63.53), whereas in those not using biologic treatment it was 67.41 ± 72.20; median 43.37 (0.41–262.45); however, the difference was not statistically significant (p = 0.232). Similarly, no significant difference in CALLY values was found according to anakinra use (p = 0.122), canakinumab use (p = 0.989), or the presence of amyloidosis (p = 0.396) (Table 8).

When variables related to colchicine treatment were examined, CALLY values were similar according to locally manufactured colchicine use (p = 0.686) and imported colchicine use (p = 0.088). In contrast, in patients with ineffective locally manufactured colchicine treatment, the CALLY value was 47.03 ± 71.39; median 11.15 (0.49–238.74), whereas in patients responding to colchicine it was 72.52 ± 70.02; median 49.08 (0.41–262.45), and the difference was statistically significant (p = 0.007) (Table 8).

In addition, no significant difference was found in CALLY values according to the presence of ankylosing spondylitis (p = 0.378) and the presence of other mutations (p = 0.329). Since only one patient was receiving anti-TNF treatment, no statistical comparison was performed for this variable. These findings indicate that the CALLY index is particularly associated with the presence of attacks, disease activity, and response to colchicine treatment, whereas it does not show significant differences according to other clinical features (Table 8).

In the FMF group, the AGE-SAF levels were higher in patients with the presence of comorbidity (2.24 ± 0.55; 2.20 [1.40–3.60]) than in those without comorbidity, in whom it was 1.98 ± 0.36; 2.00 [1.30–3.00], and the difference was significant (p = 0.035). Similarly, AGE-SAF levels were higher in patients with hypertension (p = 0.002). AGE-SAF levels were also higher in the presence of atherosclerosis, and the difference was significant (p = 0.027). In contrast, no significant difference in AGE-SAF levels was found according to coronary artery disease (p = 0.103) and other comorbidities (p = 0.245) (Table 9).

Table 9.

Comparison of AGE-SAF and CALLY according to comorbidities.

Variable	AGE-SAF No	AGE-SAF Yes	p	CALLY No	CALLY Yes	p
Presence of comorbidity	1.98 ± 0.36; 2.00 (1.30–3.00)	2.24 ± 0.55; 2.20 (1.40–3.60)	0.035	75.18 ± 79.51; 46.69 (0.41–262.45)	43.02 ± 39.60; 36.47 (3.92–169.46)	0.298
Hypertension	1.99 ± 0.39; 1.95 (1.30–3.00)	2.41 ± 0.54; 2.20 (1.70–3.60)	0.002	70.12 ± 74.73; 45.15 (0.41–262.45)	41.59 ± 43.40; 35.31 (3.93–169.46)	0.354
Coronary artery disease	2.03 ± 0.40; 2.00 (1.30–3.01)	2.52 ± 0.75; 2.20 (1.80–3.60)	0.103	66.69 ± 71.64; 45.08 (0.41–262.45)	45.11 ± 62.95; 22.38 (4.28–169.46)	0.370
Atherosclerosis	2.03 ± 0.40; 2.00 (1.30–3.01)	2.80 ± 0.77; 2.80 (2.00–3.60)	0.027	65.71 ± 71.09; 43.46 (0.41–262.45)	54.62 ± 77.75; 22.38 (4.28–169.46)	0.550
Other comorbidity	2.02 ± 0.37; 2.00 (1.30–3.00)	2.23 ± 0.63; 2.15 (1.40–3.60)	0.245	71.65 ± 76.67; 46.46 (0.41–262.45)	40.48 ± 33.77; 39.88 (3.92–113.81)	0.376

AGE-SAF, advanced glycation end products measured by skin autofluorescence; CALLY, C-reactive protein–albumin–lymphocyte index.

In terms of CALLY, no statistically significant difference was found between the groups with respect to the presence of comorbidity, hypertension, coronary artery disease, atherosclerosis, and other comorbidities (all p > 0.05) (Table 9).

When AGE-SAF levels were evaluated according to the presence of genetic mutations in the FMF group, no statistically significant difference in AGE-SAF levels was found for most mutations (p > 0.05). However, in some mutations, AGE-SAF levels were numerically higher. In patients with the compound M694V/E148Q mutation, AGE-SAF values were 2.48 ± 0.68; median 2.25 (1.60–3.60), whereas in patients without this mutation it was 2.02 ± 0.39; median 2.00 (1.30–3.01), and AGE-SAF levels were higher in patients with the mutation (p = 0.037). Similarly, AGE-SAF levels were higher in patients with the R761H heterozygous mutation than in those without the mutation (2.66 ± 0.66; median 2.80 [1.80–3.40] vs. 2.03 ± 0.42; median 2.00 [1.30–3.60], respectively; p = 0.011) (Table 10).

Table 10.

Comparison of AGE-SAF and CALLY according to genetic mutations.

Genetic mutation	AGE-SAF non-carriers	AGE-SAF carriers	p	CALLY Non-carriers	CALLY carriers	p
M694V heterozygous	2.09 ± 0.48; 2.00 (1.30–3.60)	2.04 ± 0.42; 2.00 (1.40–3.30)	0.578	76.18 ± 77.25; 60.62 (0.49–262.45)	57.79 ± 65.93; 37.16 (0.41–241.86)	0.322
R202Q heterozygous	2.08 ± 0.47; 2.00 (1.30–3.60)	2.03 ± 0.39; 2.00 (1.40–3.00)	0.599	67.03 ± 72.33; 43.37 (0.41–262.45)	62.40 ± 69.58; 40.08 (0.49–238.74)	0.761
V726A heterozygous	2.05 ± 0.42; 2.00 (1.30–3.60)	2.19 ± 0.58; 2.05 (1.60–3.60)	0.276	63.30 ± 67.34; 43.29 (0.41–262.45)	79.84 ± 98.29; 39.06 (3.93–241.86)	0.736
E148Q heterozygous	2.05 ± 0.42; 2.00 (1.30–3.60)	2.13 ± 0.58; 2.00 (1.40–3.60)	0.875	64.96 ± 71.73; 43.29 (0.41–262.45)	66.55 ± 69.12; 42.95 (0.49–238.74)	0.892
m680I homozygous	2.06 ± 0.45; 2.00 (1.30–3.60)	2.27 ± 0.40; 2.20 (1.90–2.70)	0.332	66.76 ± 72.12; 43.46 (0.41–262.45)	20.65 ± 17.47; 12.56 (6.81–42.58)	0.385
M694V homozygous	2.03 ± 0.42; 2.00 (1.30–3.60)	2.32 ± 0.61; 2.20 (1.40–3.60)	0.287	71.75 ± 75.54; 46.93 (0.41–262.45)	8.47 ± 8.09; 4.29 (0.49–22.47)	0.007
R202Q homozygous	2.05 ± 0.44; 2.00 (1.30–3.60)	2.18 ± 0.50; 2.10 (1.40–3.00)	0.658	69.87 ± 73.93; 45.22 (0.41–262.45)	11.17 ± 13.86; 4.29 (0.49–36.35)	0.027
m680I heterozygous	2.03 ± 0.42; 2.00 (1.30–3.60)	2.28 ± 0.57; 2.10 (1.50–3.60)	0.360	70.28 ± 72.65; 45.22 (0.41–262.45)	33.43 ± 49.51; 12.56 (0.49–169.46)	0.526
Compound M694V/V726A	2.06 ± 0.45; 2.00 (1.30–3.60)	2.16 ± 0.35; 2.20 (1.70–2.60)	0.558	66.61 ± 72.37; 43.29 (0.41–262.45)	42.04 ± 43.38; 18.66 (3.93–113.81)	0.417
Compound M694V/M680I	2.05 ± 0.43; 2.00 (1.30–3.60)	2.34 ± 0.61; 2.20 (1.50–3.20)	0.070	67.09 ± 71.57; 45.08 (0.41–262.45)	33.99 ± 71.66; 4.29 (0.49–163.44)	0.460
Compound M694V/E148Q	2.02 ± 0.39; 2.00 (1.30–3.01)	2.48 ± 0.68; 2.25 (1.60–3.60)	0.037	69.11 ± 72.69; 45.08 (0.41–262.45)	26.56 ± 52.47; 1.84 (0.49–153.13)	0.148
M694V homozygous + R202Q homozygous	2.07 ± 0.45; 2.00 (1.30–3.60)	1.93 ± 0.29; 1.85 (1.70–2.30)	0.652	68.35 ± 72.61; 45.22 (0.41–262.45)	0.98 ± 0.81; 0.80 (0.49–1.84)	0.017
Compound M694V/R202Q	2.03 ± 0.42; 2.00 (1.30–3.60)	2.13 ± 0.48; 2.05 (1.40–3.60)	0.366	69.67 ± 75.43; 43.37 (0.41–262.45)	56.70 ± 61.43; 39.96 (0.49–238.74)	0.724
R761H heterozygous	2.03 ± 0.42; 2.00 (1.30–3.60)	2.66 ± 0.66; 2.80 (1.80–3.40)	0.011	67.59 ± 71.72; 43.46 (0.41–262.45)	26.12 ± 40.47; 1.84 (0.49–94.91)	0.225
Other mutation	2.04 ± 0.39; 2.00 (1.30–3.01)	2.30 ± 0.78; 2.15 (1.40–3.60)	0.461	60.19 ± 65.81; 42.77 (0.41–262.45)	114.63 ± 102.60; 120.46 (0.49–241.86)	0.329

Mutation groups with insufficient sample size (K361S heterozygous, V726A homozygous, Compound M680I/E148Q, Compound M680I/V726A, Compound M694V/R761H heterozygous, and R761H homozygous) were excluded from formal comparative analyses. Mutation-positive and mutation-negative groups indicate the presence or absence of the specified genetic variant, respectively. Data are presented as mean ± standard deviation and median (minimum–maximum). *Insufficient sample size.

Although AGE-SAF values were numerically higher in some mutations, they did not reach statistical significance. For example, AGE-SAF levels were higher in patients with the compound M694V/M680I mutation than in those without the mutation (2.34 ± 0.61 vs. 2.05 ± 0.43), although the difference did not reach statistical significance (p = 0.070). In contrast, in some mutations the AGE-SAF value appeared lower, but this was not significant. For example, in patients with the M694V homozygous + R202Q homozygous mutation, the AGE-SAF value was 1.93 ± 0.29, whereas in patients without the mutation it was 2.07 ± 0.45, and the difference was not significant (p = 0.652). Overall, AGE-SAF levels did not differ significantly across most MEFV mutation groups. However, significantly higher AGE-SAF levels were observed in patients carrying the compound M694V/E148Q mutation and the R761H heterozygous mutation. Nevertheless, these findings should be interpreted cautiously because of the small sample sizes of the mutation subgroups (Table 10).

When evaluated in terms of the CALLY index, marked differences were found in some mutations. In patients with the M694V homozygous mutation, the CALLY value was 8.47 ± 8.09; median 4.29 (0.49–22.47), whereas in patients without this mutation it was 71.75 ± 75.54; median 46.93 (0.41–262.45), indicating that the CALLY value was markedly lower in the presence of the mutation (p = 0.007). Similarly, in patients with the R202Q homozygous mutation, the CALLY value was 11.17 ± 13.86; median 4.29 (0.49–36.35), whereas in patients without the mutation it was 69.87 ± 73.93; median 45.22 (0.41–262.45), again showing that the CALLY value was lower (p = 0.027). In addition, in patients with the M694V homozygous + R202Q homozygous mutation combination, the CALLY value was 0.98 ± 0.81; median 0.80 (0.49–1.84), whereas in those without this combination it was 68.35 ± 72.61; median 45.22 (0.41–262.45), indicating that the CALLY value was much lower in the presence of the mutation (p = 0.017). No significant difference in CALLY values was found for most of the other mutations (p > 0.05). However, numerical differences were observed in some mutations. For example, in patients with other mutations, the CALLY value was 114.63 ± 102.60, whereas in those without mutations it was 60.19 ± 65.81, and although the values appeared higher, the difference was not significant (p = 0.329). Overall, these findings indicate that especially M694V and R202Q homozygous mutations are associated with a marked decrease in the CALLY index, whereas heterozygous mutations and most mutation combinations do not have a significant effect on CALLY (Table 10).

A linear regression analysis was performed to evaluate the determinants of AGE-SAF levels in the FMF group. In the univariate analysis, a negative relationship was observed between AGE-SAF and log(CALLY) (β=−0.083; p = 0.037). In addition, age and the R761H heterozygous mutation were positively associated with AGE-SAF (p < 0.001 and p = 0.006, respectively). The compound M694V/E148Q mutation was also associated with AGE-SAF (p = 0.034). In the multivariate model, age was independently associated with higher AGE-SAF levels (β = 0.012, 95% CI: 0.004–0.020, p = 0.003). In addition, the R761H heterozygous mutation was independently associated with higher AGE-SAF levels (β = 0.525, 95% CI: 0.050–1.000, p = 0.030); however, this finding should be interpreted with caution given the limited sample size. In contrast, log(CALLY), BMI, HbA1c, TG/HDL ratio, number of comorbidities, and number of attacks in the last three months were not significant in the multivariate model. The model was statistically significant and explained approximately 54% of the variance in AGE-SAF (R² = 0.537), indicating a moderate explanatory capacity of the model (Table 11). The final model explained 53.7% of the variance in AGE-SAF (R² = 0.537). Given the sample size and the exploratory nature of some covariates included in the analysis, these findings should be interpreted cautiously. In addition, given the relatively modest sample size and the presence of small mutation subgroups, the regression model may be susceptible to overfitting. Therefore, these regression findings should be considered exploratory and require confirmation in larger cohorts.

Table 11.

Regression analyses for AGE-SAF.

Variable	Univariable β (95% CI)	p	Multivariable β (95% CI)	p
log(CALLY)	− 0.083 (− 0.161 to − 0.005)	0.037	− 0.062 (− 0.141 to 0.016)	0.121
Age	0.015 (0.007 to 0.024)	< 0.001	0.012 (0.004 to 0.020)	0.003
Sex (male)	− 0.112 (− 0.268 to 0.044)	0.158	− 0.114 (− 0.270 to 0.042)	0.153
BMI	0.004 (− 0.012 to 0.020)	0.626	0.001 (− 0.012 to 0.014)	0.881
HbA1c (%)	0.136 (− 0.092 to 0.365)	0.240	0.091 (− 0.144 to 0.325)	0.449
TG/HDL ratio	0.022 (− 0.010 to 0.055)	0.176	0.026 (− 0.012 to 0.065)	0.178
Number of comorbidities	0.089 (− 0.056 to 0.235)	0.224	0.024 (− 0.109 to 0.156)	0.728
Number of attacks in last 3 months	0.056 (− 0.006 to 0.118)	0.077	0.026 (− 0.042 to 0.094)	0.455
Compound M694V/E148Q	0.310 (0.025 to 0.595)	0.034	0.231 (− 0.047 to 0.510)	0.104
R761H heterozygous	0.642 (0.184 to 1.100)	0.006	0.525 (0.050 to 1.000)	0.030

Model summary (multivariate): R² = 0.537 Adjusted R² = 0.476 Model p = 7.76 × 10⁻⁷.

Discussion

In this study, we evaluated tissue accumulation of AGEs measured by SAF in patients with FMF and investigated its relationship with the CALLY index, inflammatory markers, metabolic parameters, clinical characteristics, comorbidity burden, and MEFV gene mutations. The principal findings were that AGE-SAF levels were significantly higher in patients with FMF than in healthy controls and that increased AGE-SAF levels were associated with inflammatory activity, metabolic disturbances, attack frequency, and comorbidity burden. In addition, a significant inverse relationship was observed between AGE-SAF and the CALLY index. Taken together, these findings may suggest that tissue AGE accumulation is associated with the combined inflammatory and metabolic burden in FMF. However, given the cross-sectional design and the absence of formal correction for multiple comparisons, these findings should be interpreted with caution, and no causal inferences can be made.

From a clinical perspective, increased tissue AGE accumulation likely reflects prolonged exposure to chronic inflammation and metabolic stress. Unlike circulating inflammatory markers, which may fluctuate during acute inflammatory episodes, SAF provides a more stable estimate of cumulative glycation burden and long-term cardiometabolic stress^{6,7,11,13,14,18}. Prior studies have shown that higher SAF is associated with endothelial dysfunction, arterial stiffness, and major adverse cardiovascular events^11,13–18, with prospective evidence linking SAF to mortality²⁹, supporting the clinical relevance of even modest elevations in tissue AGE accumulation. In addition, recent data suggest that composite inflammatory indices, including the CALLY index, reflect inflammatory burden in FMF³⁰.

Therefore, SAF may capture subclinical cumulative damage that is not fully reflected by conventional inflammatory biomarkers.

FMF is increasingly recognized as a condition characterized not only by episodic inflammatory attacks but also by persistent subclinical inflammation^1–5. This chronic inflammatory milieu promotes oxidative stress and metabolic dysregulation, both of which contribute to AGE formation. AGEs are generated through non-enzymatic glycation and oxidation of proteins and lipids and accumulate progressively in tissues under sustained inflammatory conditions^6,7. In addition to being passive by-products, AGEs exert biological effects through interaction with the RAGE, leading to activation of proinflammatory signaling pathways and amplification of oxidative stress^8–12. This mechanism suggests a self-perpetuating cycle in which chronic inflammation promotes AGE formation, and AGE accumulation further sustains inflammatory signaling. The observed positive correlation between AGE-SAF and CRP in our study supports this concept and may indicate that tissue AGE accumulation reflects overall inflammatory burden in FMF.

Chronic inflammation in FMF is also associated with long-term complications, including endothelial dysfunction, cardiovascular disease, and AA amyloidosis^{1–5,31–33}. Although amyloidosis was not directly evaluated in the present study, AGE accumulation—reflecting prolonged exposure to inflammatory and oxidative stress—may be linked to pathways involved in long-term tissue injury. However, this interpretation remains speculative and should not be overextended without dedicated mechanistic or longitudinal studies.

Consistent with previous studies in chronic inflammatory conditions, AGE-SAF levels were associated with both inflammatory and metabolic parameters, including CRP, HbA1c, triglycerides, total cholesterol, and the TG/HDL ratio^6,7,9,11. In addition, AGE-SAF correlated with attack frequency and active disease status, suggesting that recurrent inflammatory episodes and persistent disease activity contribute to progressive AGE accumulation. Notably, the association between AGE-SAF and the metabolic burden score was observed in the FMF group but not in controls, supporting the concept that AGE accumulation reflects the combined effects of inflammation and metabolic dysregulation in FMF rather than metabolic exposure alone.

Another important finding was the relationship between AGE-SAF and comorbidity burden. Higher AGE-SAF levels were observed in patients with cardiovascular comorbidities, particularly hypertension and atherosclerosis. This finding is biologically plausible, as AGEs contribute to endothelial dysfunction and vascular injury through oxidative stress, reduced nitric oxide bioavailability, matrix crosslinking, and vascular stiffening^9,11,12,34. Accordingly, our findings further support the potential role of AGE accumulation as a marker of long-term cardiovascular risk in FMF.

A key observation of this study was the inverse relationship between AGE-SAF and the CALLY index. The CALLY index, calculated from CRP, albumin, and lymphocyte count, has emerged as a composite biomarker integrating inflammatory activity, immune status, and nutritional reserve^20–22. Although it was initially developed in oncological settings, recent evidence suggests that the CALLY index may also be informative in rheumatologic and chronic inflammatory disorders, including FMF^26–28,30. In rheumatoid arthritis, a recent retrospective cohort study reported that lower CALLY values were associated with higher all-cause mortality, suggesting that CALLY may capture clinically meaningful aspects of systemic inflammatory and immune burden²⁶. In addition, a cross-sectional study in rheumatoid arthritis showed that immune-nutrition indices were associated with disease activity, supporting the broader relevance of composite inflammatory–nutritional markers in rheumatology²⁷. More recently, a study in primary Sjögren’s syndrome demonstrated that both CALLY and HALP were negatively associated with disease activity, extending the possible utility of such biomarkers to other systemic rheumatic diseases²⁸.

In this context, the inverse association observed in our study suggests that AGE-SAF and the CALLY index capture distinct but complementary aspects of disease burden in FMF. AGE-SAF likely reflects cumulative exposure to inflammatory and metabolic stress over time, whereas the CALLY index appears to reflect more dynamic changes related to current inflammatory and nutritional–immune status. This distinction may be clinically relevant, because FMF is characterized by both recurrent overt attacks and ongoing subclinical inflammation. The combined assessment of AGE-SAF and CALLY may therefore offer a broader view of disease burden by integrating cumulative tissue injury with more immediate systemic inflammatory status. At the same time, interpretation of CALLY-based associations requires caution, because CRP is a direct component of the formula. Accordingly, part of the strong inverse relationship between CALLY and inflammatory variables may reflect mathematical coupling rather than a fully independent biological relationship. This limitation does not negate the potential clinical value of the index, but it does argue for cautious interpretation and for validation against hard clinical outcomes.

The genetic findings should be interpreted with caution. Higher AGE-SAF levels were observed in patients carrying certain MEFV mutations, particularly the R761H heterozygous and M694V/E148Q compound mutations. In parallel, markedly lower CALLY index values were observed in patients with high-penetrance mutations, especially M694V homozygous and R202Q homozygous variants, suggesting a higher inflammatory burden in these subgroups. Notably, M694V homozygosity and compound mutations such as M694V/E148Q have been associated with a more severe disease phenotype and an increased risk of AA amyloidosis^2,3,31,32. In this context, the observed pattern of higher AGE-SAF and lower CALLY values in these mutation groups may suggest that these markers could reflect an underlying predisposition to more severe inflammatory and metabolic burden. However, several mutation subgroups were small, and the analyses were not powered for robust subgroup comparisons. Therefore, these findings should be considered exploratory and hypothesis-generating, and it remains uncertain whether AGE-SAF and the CALLY index can reliably predict genotype-associated complications such as amyloidosis without validation in larger, longitudinal cohorts. Because the CALLY index is derived from CRP, albumin, and lymphocyte count, extreme values of these parameters may disproportionately influence the resulting score. In particular, very low CRP concentrations may lead to markedly elevated CALLY values, whereas substantially elevated CRP levels may result in very low CALLY values. Therefore, interpretation of the CALLY index should consider the underlying laboratory components rather than relying solely on the composite value. Future studies should establish clinically relevant reference ranges and evaluate the robustness of the CALLY index under extreme laboratory conditions. In the present cohort, CALLY values were calculated using albumin concentrations ranging from 36.1 to 54.6 g/L, lymphocyte counts ranging from 0.39 to 3.95 × 10³/µL, and CRP concentrations ranging from 0.5 to 208 mg/L. Interpretation of CALLY values beyond these observed ranges should therefore be made with caution until further validation studies are available.

Several limitations should be acknowledged. The cross-sectional design precludes causal inference between AGE accumulation and inflammatory or metabolic processes, and the single center setting with a relatively modest sample size may limit generalizability. Multiple subgroup and correlation analyses were performed without formal correction for multiple comparisons, increasing the risk of type I error; therefore, some of the observed associations may represent chance findings rather than true biological relationships. In addition, some composite indices used in this study, including the inflammatory burden score and metabolic burden score, were developed for exploratory purposes and have not been externally validated; accordingly, they should not be interpreted as clinically established indices, and their clinical applicability remains uncertain. Although age was adjusted for in regression analyses, residual confounding related to age cannot be completely excluded, particularly because the control group was older than the FMF group and not age-matched. Given the relatively limited sample size and the number of predictors included in the regression model, the possibility of model overfitting cannot be excluded. In addition, genetic subgroup analyses were based on relatively small sample sizes, which limits the robustness of mutation-specific findings; therefore, these results should be considered exploratory and hypothesis-generating. Furthermore, SAF provides an overall estimate of tissue AGE accumulation but does not distinguish between specific AGE subtypes or circulating AGE concentrations. Finally, lifestyle factors such as smoking, dietary AGE intake, and physical activity were not systematically assessed and may have influenced AGE levels. In addition, the strong inverse association between the CALLY index and CRP should be interpreted with caution, as CRP is a component of the CALLY formula, which may introduce mathematical coupling.

Despite these limitations, this study provides an integrated evaluation of tissue AGE accumulation in FMF by combining inflammatory, metabolic, clinical, and genetic data. The findings suggest that SAF may serve as a non-invasive marker of cumulative inflammatory–metabolic burden, while the CALLY index may provide complementary information regarding current inflammatory and nutritional–immune status. Together, these measures may improve characterization of disease burden in FMF, particularly in patients with persistent subclinical inflammation and increased cardiometabolic risk.

Conclusions

Tissue AGE accumulation assessed by skin autofluorescence is increased in patients with FMF and is associated with inflammatory activity, metabolic disturbances, and overall disease burden. The observed inverse relationship between AGE-SAF and the CALLY index suggests that these markers provide complementary insights into cumulative inflammatory–metabolic stress. The combined use of SAF and composite inflammatory indices may improve disease assessment and risk stratification in FMF. Further prospective studies are required to validate these findings and to clarify their clinical implications.

Abbreviations

FMF

Familial Mediterranean fever

AGEs

Advanced glycation end products

SAF

Skin autofluorescence

AGE-SAF

Skin autofluorescence-derived advanced glycation end products

CALLY

C-reactive protein–albumin–lymphocyte index

CRP

C-reactive protein

ESR

Erythrocyte sedimentation rate

SAA

Serum amyloid A

BMI

Body mass index

HbA1c

Glycated hemoglobin

LDL

Low-density lipoprotein

HDL

High-density lipoprotein

TG

Triglyceride

SII

Systemic immune-inflammation index

HALP

Hemoglobin, albumin, lymphocyte, platelet index

IBS

Inflammatory burden score

Author contributions

A.G. conceived and designed the study. A.G. and S.T.G. collected the data. A.G. performed the statistical analyses. A.G. drafted the manuscript. S.T.G. critically revised the manuscript for important intellectual content. All authors approved the final version of the manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Informed consent

Written informed consent was obtained from all participants prior to their inclusion in the study.

Institutional review board statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Bursa City Hospital (Approval No: 2026-3/8; Date: 4 February 2026). The study was registered at ClinicalTrials.gov (Identifier: NCT07439341).

Competing interests

The authors declare no competing interests.