Serum FAM19A5 in newly diagnosed type 2 diabetes mellitus and its association with metabolic and inflammatory parameters

Abstract

We aimed to investigate serum levels of FAM19A5, a novel adipokine implicated in inflammation and metabolic regulation, and to examine its association with key metabolic and inflammatory markers in individuals newly diagnosed with type 2 diabetes mellitus (T2DM). This case-control study included 54 patients with newly diagnosed T2DM and 54 age- and sex-matched normoglycemic controls aged 35 to 65 years. Continuous variables were summarized according to their distributions and compared using the independent-samples t test or Mann–Whitney U test, as appropriate. Spearman rank correlation was used to examine associations in the pooled sample, with false discovery rate correction for multiple testing. Subgroup analyses were conducted by metabolic syndrome, insulin resistance, and body mass index (BMI) category. The incremental value of FAM19A5 beyond conventional markers was explored using multivariable logistic regression and DeLong comparisons of the areas under the curve (AUCs). Serum FAM19A5 was lower in the T2DM group than in controls (376.1 [312.1–577.6] vs 554.5 [365.6–1524.2] ng/L; P < .001). In the pooled sample, FAM19A5 was inversely correlated with fasting glucose (ρ = −0.266, P = .005), 2-hour oral glucose tolerance test (ρ = −0.308, P = .001), and HbA1c (ρ = −0.212, P = .028), but not with BMI, waist circumference, lipid parameters, homeostasis model assessment of insulin resistance (HOMA-IR), or high-sensitivity C-reactive protein (hs-CRP). After false discovery rate correction, the associations with fasting and 2-hour oral glucose tolerance test remained significant. Within the T2DM group, FAM19A5 was lower in patients with metabolic syndrome (P = .004), insulin resistance (HOMA-IR > 2.71; P = .023), or BMI ≥25 kg/m² (P = .021). FAM19A5 alone showed poor discrimination for T2DM (AUC: 0.684; sensitivity: 66.7%; specificity: 53.7%). Adding FAM19A5 to a model containing BMI, HOMA-IR, and hs-CRP improved model fit (likelihood-ratio χ² = 13.1, P < .001) but not discrimination (AUC: 0.920 vs 0.943; DeLong P = .154). Serum FAM19A5 levels were lower in newly diagnosed T2DM and remained independently associated with diabetes status after adjustment for BMI, HOMA-IR, and hs-CRP. Correlations with glycemic measures were weak, and FAM19A5 alone had poor discriminatory performance. These findings support further investigation of FAM19A5 as a molecule of potential mechanistic interest, but not as a clinically useful diagnostic biomarker. Larger longitudinal studies are needed.

1. Introduction

Diabetes mellitus is a growing global health concern. It currently affects over 537 million adults worldwide and is estimated to reach 783 million by 2045.^[1] The prevalence of diabetes mellitus has reached 13.2% in Turkey, and the follow-up, treatment, and prevention of complications of this chronic disease constitute a significant burden, especially in primary care.^[2]

Type 2 diabetes mellitus (T2DM) is the most common form of diabetes, characterized by insulin resistance, β-cell dysfunction, low-grade chronic inflammation, and disrupted adipokine signaling.^[3] Adipokines are bioactive molecules secreted by the adipose tissue and play a central role in the metabolic and inflammatory axes. They are known to regulate glucose homeostasis, lipid metabolism, and insulin sensitivity.^[4,5] Persistent inflammation contributes to the progression of T2DM, as pro-inflammatory mediators such as interleukin-6, tumor necrosis factor-alpha, and C-reactive protein (CRP) have been linked to glucotoxicity, lipotoxicity, and insulin resistance.^[6,7] These inflammatory responses are also associated with cardiovascular complications, including atherosclerosis and endothelial dysfunction.^[8]

FAM19A5, a relatively novel adipokine encoded by the TAFA1-TAFA5 gene family, has recently gained attention for its potential role in metabolic disorders. Beyond its known functions in inflammation, cellular signaling, angiogenesis, and tissue homeostasis,^[9,10] FAM19A5 has been implicated in regulating vascular smooth muscle cell proliferation and maintaining cardiovascular stability.^[11,12] Reduced FAM19A5 levels have been reported in individuals with obesity, metabolic syndrome (MetS), and coronary artery disease, suggesting its involvement in metabolic dysregulation.^[13,14] Additionally, emerging research points to a possible role of FAM19A5 in neuroinflammatory and neurodegenerative processes and in thyroid cancer prognosis.^[10,15]

There is increasing evidence of a relationship between FAM19A5 and T2DM, especially in newly diagnosed patients. However, this aspect remains poorly understood. Given the chronic and progressive nature of T2DM, early detection and monitoring are critical to mitigate long-term complications.^[16] Family physicians also play an important role in patient management. Thus, identification of novel biomarkers may support earlier risk stratification and personalized interventions.

In light of these considerations, this study aimed to investigate serum FAM19A5 levels in individuals newly diagnosed with T2DM and assess its associations with key metabolic and inflammatory markers. A secondary, explicitly exploratory objective was to quantify whether FAM19A5 adds discriminatory information beyond conventional metabolic and inflammatory markers.

2. Methods

2.1. Study design

This case-control study was conducted between August 2023 and April 2024. All variables were measured at a single time point, with no longitudinal follow-up. The study included 54 patients with newly diagnosed T2DM and 54 age- and sex-matched normoglycemic controls aged 35 to 65 years. Participants were recruited from the Internal Medicine and Family Medicine outpatient clinics of Bozyaka Training and Research Hospital (Fig. 1).

Figure 1.

Flow diagram of participant selection.

This study was designed to compare serum FAM19A5 levels and related metabolic markers between the groups to provide initial insights for future longitudinal research. Nondiabetic controls were selected solely on the basis of normal glucose tolerance; no criterion of metabolic health (e.g., normal body mass index (BMI), absence of dyslipidemia, or absence of MetS) was applied. The diagnosis of T2DM was established using the American Diabetes Association 2022 criteria.^[17]

The participants in both groups were matched for age and sex. Other potential confounding variables, such as physical activity level, diet, and medication use, were not systematically controlled. Because the groups were not matched for BMI (Table 1), BMI was entered as a covariate in the multivariable logistic regression models presented in Section 3.3 (Table 3).

Table 1.

Demographic and laboratory characteristics of the study groups.

Variable	Controls (n = 54)	nT2DM (n = 54)	P	Test
Male/female, n (%)	31 (57.4)/23 (42.6)	29 (53.7)/25 (46.3)	.846	χ 2
Age, yr	43.0 (40.0–45.8)	45.0 (42.0–50.0)	.146	MWU
BMI, kg/m2	24.9 (23.7–26.7)	25.9 (24.6–28.3) ↑	.014*	MWU
Metabolic syndrome, yes/no, n (%)	39 (72.2)/15 (27.8)	43 (79.6)/11 (20.4)	.368	χ2
Waist circumference, cm	105.1 ± 6.0	116.5 ± 8.7 ↑	<.001*	t test
Fasting glucose, mg/dL	89.5 ± 9.0	135.6 ± 34.2 ↑	<.001*	t test
2-h OGTT glucose, mg/dL	113.6 ± 22.5	218.8 ± 55.8 ↑	<.001*	t test
HbA1c, %	5.4 ± 0.36	7.4 ± 0.87 ↑	<.001*	t test
Insulin, µIU/mL	10.4 (7.3–15.2)	16.3 (11.6–24.4) ↑	.002*	MWU
HOMA-IR	2.3 (1.5–3.7)	5.3 (3.4–8.3) ↑	<.001*	MWU
Total cholesterol, mg/dL	195.0 ± 35.1	231.5 ± 39.9 ↑	<.001*	t test
LDL-C, mg/dL	118.7 ± 23.7	138.4 ± 46.4 ↑	.007*	t test
HDL-C, mg/dL	50.4 ± 13.3	42.5 ± 13.5 ↓	.003*	t test
Triglycerides, mg/dL	136.9 ± 67.8	157.2 ± 87.6	.173	t test
BUN, mg/dL	26.2 ± 5.9	26.7 ± 7.6	.695	t test
Uric acid, mg/dL	5.1 ± 1.3	6.3 ± 1.2 ↑	<.001*	t test
Creatinine, mg/dL	0.79 ± 0.15	0.77 ± 0.15	.327	t test
AST, IU/L	17.6 (15.0–23.8)	16.1 (14.0–21.5)	.401	MWU
ALT, IU/L	18.5 (14.0–23.3)	21.4 (15.3–30.0)	.165	MWU
LDH, IU/L	159.8 ± 25.9	185.8 ± 31.6 ↑	<.001*	t test
Hemoglobin, g/dL	14.5 (13.3–15.4)	13.9 (13.2–15.1)	.740	MWU
Platelet, x103/uL	271.0 (223.0–294.0)	282.0 (248.2–333.5) ↑	.032*	MWU
RBC, x106/uL	4.9 (4.5–5.1)	4.9 (4.7–5.1)	.305	MWU
WBC, x103/uL	6.9 ± 1.2	8.8 ± 2.3 ↑	<.001*	t test
hs-CRP, mg/L	1.7 (0.9–2.6)	4.2 (1.5–11.7) ↑	<.001*	MWU
FAM19A5, ng/L	554.5 (365.6–1524.2)	376.1 (312.1–577.6) ↓	<.001*	MWU

Data are presented as median (interquartile range) for non-normally distributed variables and mean ± standard deviation for normally distributed variables. The test used for each variable is shown in the “Test” column. MWU, Mann–Whitney U test. Normality was assessed using the Shapiro–Wilk test, and categorical variables were compared using Pearson’s chi-squared test. ↑ and ↓ indicate values significantly higher and lower, respectively, in the nT2DM group than in controls.

^*P < 0.05. Metabolic syndrome was defined according to the NCEP ATP III criteria.

2-h OGTT = 2-h oral glucose tolerance test, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, BUN = blood urea nitrogen, HbA1c = glycosylated hemoglobin, HDL-C = high-density lipoprotein cholesterol, HOMA-IR = homeostasis model assessment of insulin resistance, hs-CRP = high-sensitivity C-reactive protein, LDH = lactate dehydrogenase, LDL-C = low-density lipoprotein cholesterol, nT2DM = newly diagnosed type 2 diabetes mellitus, RBC = red blood cell, WBC = white blood cell.

Table 3.

Logistic regression models for newly diagnosed T2DM versus normoglycemic controls.

Model	Variable	OR (95% CI)	P	Model AUC (95% CI)
Model 1: FAM19A5 alone	FAM19A5 (per 100 ng/L)	0.873 (0.801–0.950)	.002	0.684 (0.584–0.784)
Model 2: BMI + HOMA-IR + hs-CRP	BMI	1.544 (1.096–2.175)	.013	0.920 (0.865–0.974)
Model 2: BMI + HOMA-IR + hs-CRP	HOMA-IR	2.125 (1.542–2.929)	<.001	0.920 (0.865–0.974)
Model 2: BMI + HOMA-IR + hs-CRP	hs-CRP	1.312 (1.114–1.545)	.001	0.920 (0.865–0.974)
Model 3: Model 2 + FAM19A5	BMI	1.803 (1.195–2.719)	.005	0.943 (0.897–0.989)
Model 3: Model 2 + FAM19A5	HOMA-IR	2.369 (1.607–3.492)	<.001	0.943 (0.897–0.989)
Model 3: Model 2 + FAM19A5	hs-CRP	1.292 (1.087–1.536)	.004	0.943 (0.897–0.989)
Model 3: Model 2 + FAM19A5	FAM19A5 (per 100 ng/L)	0.817 (0.717–0.932)	.003	0.943 (0.897–0.989)

ORs for FAM19A5 are reported per 100 ng/L increase. Fasting glucose, 2-h OGTT glucose, and HbA1c were excluded because they defined group membership and produced complete separation. DeLong test: Model 1 vs Model 2, P < .001; Model 2 vs Model 3, P = .154. Likelihood-ratio test for adding FAM19A5 to Model 2: χ² = 13.11, 1 df, P < .001. The models were exploratory and were not internally or externally validated.

AUC = area under the receiver operating characteristic curve, BMI = body mass index, CI = confidence interval, FAM19A5 = family with sequence similarity 19 member A5, HOMA-IR = homeostasis model assessment of insulin resistance, hs-CRP = high-sensitivity C-reactive protein, OR = odds ratio, T2DM = type 2 diabetes mellitus.

2.2. Diagnostic criteria for T2DM (American Diabetes Association 2022)

• Fasting plasma glucose (FPG)≥126 mg/dL

• 2-hour plasma glucose ≥200 mg/dL after a 75 g oral glucose tolerance test

• Glycosylated hemoglobin (HbA1c)≥6.5%

2.3. Criteria for nondiabetic (normoglycemic) controls

• FPG <100 mg/dL

• 2-hour plasma glucose <140 mg/dL

• HbA1c <5.6%

All participants were categorized as having or not having MetS using the National Cholesterol Education Program Adult Treatment Panel III criteria.^[18,19] MetS was defined as the presence of at least 3 of the following:

1. Central obesity (waist circumference ≥102 cm in men and ≥88 cm in women)

2. Hypertriglyceridemia (≥ 150 mg/dL or on treatment)

3. Low high-density lipoprotein cholesterol (HDL-C) level (< 40 mg/dL for men and <50 mg/dL for women)

4. Hypertension (≥ 130/85 mm Hg or on treatment)

5. Elevated fasting glucose (≥ 100 mg/dL)

Because nondiabetic controls were required to have FPG <100 mg/dL, criterion 5 could not, by definition, be met in this group; a control classified as having MetS therefore satisfied at least 3 of the 4 remaining criteria.

2.4. Anthropometric and clinical assessments

Standardized measurements were applied to all participants, including height, weight, waist circumference, and blood pressure. BMI was calculated as weight (kg)/height (m²). Waist circumference was measured at the midpoint between the lower rib and iliac crest after normal expiration. Blood pressure and pulse were recorded after 10 minutes of rest.

2.5. Exclusion criteria

Exclusion criteria included:

• Type 1 diabetes, thyroid disorders, malignancies, or other chronic systemic diseases

• Pregnancy or breastfeeding

• History of alcohol, tobacco, or substance use

• Use of medications that affect glucose metabolism (e.g., corticosteroids, exogenous hormones)

2.6. Biochemical and FAM19A5 measurements

Venous blood samples were collected in the morning (08:00–09:00) after at least 10 hours of fasting. Biochemical analyses included glycemic parameters (fasting plasma glucose, postprandial glucose, HbA1c), lipid profile (total cholesterol, LDL-C, HDL-C, triglycerides), the inflammatory marker high-sensitivity C-reactive protein (hs-CRP), liver function tests (alanine aminotransferase, aspartate aminotransferase), renal function tests (urea, creatinine), and a complete blood count.

Serum FAM19A5 concentrations were determined using a commercial ELISA kit (Bioassay Technology Laboratory, Jiaxing, Zhejiang, China; catalog no. E6705Hu), with a sensitivity of 10.69 ng/L and intra- and inter-assay CVs <10%. The samples were centrifuged at 3000 rpm for 10 minutes at room temperature and stored at −80°C. The absorbance was measured at 450 nm using a Multiskan GO ELISA reader (Thermo Fisher Scientific, Vantaa, Finland).

2.7. Definition of insulin resistance and BMI subgroups

Insulin resistance was defined a priori as homeostasis model assessment of insulin resistance (HOMA-IR) >2.71. HOMA-IR was calculated using the method described by Matthews et al^[20] This threshold is within the range reported for adult populations, although population-based studies indicate that optimal cutoffs vary by age and sex.^[21] In our control group, 2.71 fell between the median (2.3) and third quartile (3.7) of the HOMA-IR distribution (Table 1). BMI subgroups were defined using the World Health Organization threshold for overweight (<25 vs ≥ 25 kg/m²). FAM19A5 subgroup comparisons were performed within the T2DM group and, for MetS, within the control group (Figs. 2B–D).

Figure 2.

Circulating FAM19A5 concentrations. (A) FAM19A5 levels in T2DM versus control groups. A P-value of <.05 was considered significant (*). (B) FAM19A5 levels in individuals with and without MetS in T2DM and control groups. T2DM, type 2 diabetes mellitus; MetS, metabolic syndrome. Statistical significance was set at P < .05. (C) FAM19A5 levels among T2DM patients with and without insulin resistance (HOMA-IR). A P-value of <.05 was considered significant (*). (D) Circulating FAM19A5 levels in T2DM patients based on BMI. Normal weight: BMI < 25 kg/m2; overweight or obese: BMI ≥ 25 kg/m2. A P-value of <.05 was considered significant (*). Boxes represent medians and interquartile ranges; whiskers extend to 1.5 times the IQR, and individual observations are overlaid. Sample sizes are shown below the boxes, and P values are shown above each comparison. *P <.05.

2.8. Statistical analysis

The sample size was estimated using G*Power 3.1.9.7 (Heinrich Heine University, Düsseldorf, Germany). The distribution of each continuous variable was assessed using the Shapiro–Wilk test and visual inspection of histograms and Q–Q plots. FAM19A5 was markedly right-skewed in both groups (skewness 1.38 and 2.32; Shapiro–Wilk P < .001), and log transformation did not restore normality. Non-normally distributed variables are therefore reported as medians (interquartile ranges [IQRs]) and compared using the Mann–Whitney U test. Normally distributed variables are reported as mean ± standard deviation and compared using the independent-samples t test. The test used for each variable is shown in Table 1. Categorical variables were compared using Pearson’s chi-squared test.

Associations between FAM19A5 and metabolic or inflammatory variables were assessed in the pooled sample (n = 108) using Spearman rank correlation coefficients (ρ) and corresponding P values (Table 2). Because 16 correlations were tested, P values were adjusted using the Benjamini–Hochberg false discovery rate procedure.^[22] FAM19A5 levels were also compared across subgroups defined by MetS, insulin resistance, and BMI category (Figs. 2B–D).

Table 2.

Correlations between serum FAM19A5 and metabolic and inflammatory variables in the pooled sample (n = 108).

Variable	ρ (pooled)	P
BMI	0.005	.963
Waist circumference	−0.025	.798
Fasting glucose	−0.266	.005*
2-h OGTT glucose	−0.308	.001*
HbA1c	−0.212	.028*
Insulin	−0.044	.653
HOMA-IR	−0.138	.156
Total cholesterol	−0.078	.424
LDL-C	0.091	.347
HDL-C	−0.051	.598
Triglycerides	−0.066	.496
Uric acid	−0.126	.193
LDH	0.033	.731
Platelet	−0.022	.824
WBC	−0.052	.592
hs-CRP	−0.09	.354

ρ, Spearman rank correlation coefficient.

2-h OGTT = 2-h oral glucose tolerance test, BMI = body mass index, FAM19A5 = family with sequence similarity 19 member A5, HbA1c = glycosylated hemoglobin, HDL-C = high-density lipoprotein cholesterol, HOMA-IR = homeostasis model assessment of insulin resistance, hs-CRP = high-sensitivity C-reactive protein, LDH = lactate dehydrogenase, LDL-C = low-density lipoprotein cholesterol, WBC = white blood cell.

^*P < .05 before adjustment. Benjamini–Hochberg FDR-adjusted P values were .047 for fasting glucose, .020 for 2-h OGTT glucose, and .158 for HbA1c; no other adjusted P value was < .05.

Receiver operating characteristic (ROC) analysis was used to assess the discriminatory performance of FAM19A5. The optimal cutoff was determined using the Youden index, and area under the curve (AUCs) are reported with 95% confidence intervals calculated by the Hanley–McNeil method.^[23] Three logistic regression models were compared to explore whether FAM19A5 added information beyond conventional markers (Table 3): FAM19A5 alone (Model 1); BMI, HOMA-IR, and hs-CRP (Model 2); and Model 2 plus FAM19A5 (Model 3). AUCs were compared using DeLong’s test,^[24] and the contribution of FAM19A5 was also assessed using a likelihood-ratio test. Fasting glucose, 2-h oral glucose tolerance test (2-h OGTT) glucose, and HbA1c were excluded because they defined group membership and produced complete separation (AUC = 1.000). Analyses were performed using SPSS version 26 (IBM Corp., Armonk, NY) and Python version 3.12 (SciPy, statsmodels, and scikit-learn). All tests were two-sided, with P < .05 considered statistically significant.

3. Results

3.1. Clinical and laboratory characteristics

The clinical and laboratory characteristics of the study population are summarized in Table 1. Non-normally distributed variables are presented as median (IQR) and normally distributed variables as mean ± standard deviation. FAM19A5 levels were lower in the T2DM group than in normoglycemic controls (376.1 [312.1–577.6] vs 554.5 [365.6–1524.2] ng/L; Mann–Whitney U test, P < .001; Figure 2A). The groups did not differ significantly in sex distribution (P = .846) or age (P = .146). Median BMI was higher in the T2DM group (25.9 [24.6–28.3] vs 24.9 [23.7–26.7] kg/m²; P = .014).

MetS was present in 39 controls (72.2%) and 43 patients with T2DM (79.6%; P = .368; Table 1). Because all controls had fasting glucose <100 mg/dL, those classified as having MetS met at least 3 of the other 4 criteria: central obesity, hypertriglyceridemia, low HDL-C, and hypertension. Thus, although normoglycemic, most controls were not metabolically healthy. This is addressed further in Section 4.4.

Compared with controls, patients with T2DM had higher BMI, waist circumference, fasting glucose, 2-h OGTT glucose, HbA1c, fasting insulin, HOMA-IR, total cholesterol, LDL-C, uric acid, lactate dehydrogenase, platelet count, white blood cell (WBC), and hs-CRP, and lower HDL-C (Table 1). Triglycerides, blood urea nitrogen, creatinine, aspartate aminotransferase, alanine aminotransferase, hemoglobin, and red blood cell did not differ significantly between groups. Correlations between these variables and FAM19A5 are presented in Section 3.2 and Table 2.

Subgroup analyses are shown in Figure 2B–D. Within the T2DM group, FAM19A5 levels were lower in patients with MetS than in those without it (n = 43 vs n = 11; P = .004; Figure 2B). No significant difference was observed in controls (n = 39 vs n = 15; P = .134). FAM19A5 levels were also lower in patients with insulin resistance (HOMA-IR > 2.71: n = 40; HOMA-IR ≤ 2.71: n = 14; P = .023; Figure 2C) and in those with BMI ≥ 25 kg/m² (n = 32 vs n = 22; P = .021; Figure 2D). Figure 2 displays the medians, IQRs, and distributions for these comparisons.

3.2. Correlations between FAM19A5 and metabolic and inflammatory variables

Table 2 presents the pooled sample correlations between FAM19A5 and metabolic and inflammatory variables. FAM19A5 was inversely correlated with fasting glucose (ρ = −0.266, P = .005), 2-h OGTT glucose (ρ = −0.308, P = .001), and HbA1c (ρ = −0.212, P = .028). After Benjamini–Hochberg correction, the associations with fasting glucose and 2-h OGTT glucose remained significant (adjusted P = .047 and.020, respectively), whereas the association with HbA1c did not (adjusted P = .158).

FAM19A5 was not significantly correlated with BMI (ρ = 0.005, P = .963), waist circumference (ρ = −0.025, P = .798), HOMA-IR (ρ = −0.138, P = .156), fasting insulin (ρ = −0.044, P = .653), total cholesterol (ρ = −0.078, P = .424), LDL-C (ρ = 0.091, P = .347), HDL-C (ρ = −0.051, P = .598), triglycerides (ρ = −0.066, P = .496), or hs-CRP (ρ = −0.090, P = .354). Thus, the significant pooled-sample correlations were weak and limited to glycemic measures.

3.3. ROC curve analysis and incremental value of FAM19A5

ROC analysis was used to assess the ability of FAM19A5 to distinguish T2DM from normoglycemia (Fig. 3). At the optimal cutoff of 507.9 ng/L, sensitivity was 66.7% and specificity was 53.7%. The AUC was 0.684 (95% CI, 0.584–0.784; P = .001), indicating poor discrimination. At this cutoff, 46.3% of controls were classified as positive. In this 1:1 case-control sample, the positive and negative predictive values were 59.0% and 61.7%, respectively. These findings do not support the use of FAM19A5 as a standalone diagnostic marker.

Figure 3.

ROC curves for distinguishing newly diagnosed T2DM from normoglycemic controls. Curves are shown for FAM19A5 alone (AUC 0.684), BMI, HOMA-IR, and hs-CRP (AUC 0.920), and the same model with FAM19A5 added (AUC 0.943). Glycemic variables were excluded because they defined group membership. AUCs were compared using DeLong’s test. AUC = area under the curve, BMI = body mass index.

Three logistic regression models were compared to assess whether FAM19A5 added information beyond conventional markers (Table 3). The model containing BMI, HOMA-IR, and hs-CRP had an AUC of 0.920 (95% CI, 0.865–0.974), which was higher than that of FAM19A5 alone (DeLong P < .001). Adding FAM19A5 increased the AUC to 0.943 (95% CI, 0.897–0.989). FAM19A5 remained independently associated with T2DM (odds ratio 0.82 per 100 ng/L, 95% CI, 0.72–0.93; P = .003), and model fit improved (likelihood-ratio χ² = 13.11, 1 df; P < .001). However, the increase in AUC was not significant (DeLong P = .154). HbA1c and glucose were excluded because they defined group membership. These exploratory analyses were not prespecified or internally or externally validated.

4. Discussion

T2DM is a metabolic disorder characterized by insulin resistance, chronic inflammation, and altered adipokine secretion. Recent studies have suggested that a newly identified adipokine, FAM19A5, may be involved in the development of metabolic conditions such as obesity, nonalcoholic fatty liver disease, and cardiovascular disease. However, its specific contribution to T2DM, particularly in newly diagnosed patients, remains unclear. This study examined serum FAM19A5 levels and their relationship with metabolic and inflammatory parameters in individuals with newly diagnosed T2DM.

4.1. Key findings and their implications

The main finding was that circulating FAM19A5 levels were lower in patients with newly diagnosed T2DM than in normoglycemic controls. This difference persisted after adjustment for BMI, HOMA-IR, and hs-CRP (Table 3), suggesting it was not fully explained by adiposity, insulin resistance, or inflammation.

The secondary findings should be interpreted cautiously. In the pooled sample, FAM19A5 showed weak inverse correlations with fasting glucose, 2-h OGTT glucose, and HbA1c (|ρ| ≤ 0.31). After FDR correction, only the associations with fasting and 2-h OGTT glucose remained significant. No pooled association was found with adiposity, lipid measures, insulin resistance, or inflammation. FAM19A5 was lower in T2DM subgroups with MetS, insulin resistance, or BMI ≥ 25 kg/m², but these analyses were not prespecified and remain exploratory.

The diagnostic value of FAM19A5 was limited. An AUC of 0.684 and specificity of 53.7% would result in a high false-positive rate, making FAM19A5 unsuitable for screening. Although FAM19A5 remained independently associated with T2DM and improved model fit, it did not significantly improve discrimination beyond BMI, HOMA-IR, and hs-CRP. Because T2DM can be diagnosed with inexpensive, widely available glycemic tests, FAM19A5 may be more relevant to disease biology than to diagnosis.

4.2. Comparison with previous studies

Previous studies have also linked lower FAM19A5 levels to metabolic abnormalities. Yari et al reported lower levels in patients with nonalcoholic fatty liver disease and inverse correlations with BMI, visceral fat, and liver fibrosis markers.^[25] Xie et al found lower FAM19A5 levels in children with obesity, with inverse correlations with fasting glucose, insulin, and BMI.^[26] In our pooled analysis, however, FAM19A5 was associated with glycemic measures but not with BMI or insulin resistance. This difference may reflect the relatively narrow BMI range in our sample, differences in age and clinical characteristics across studies, or variation between ELISA platforms. No international reference standard for FAM19A5 assays is currently available.

Contrasting findings have also been reported. Lee et al observed higher FAM19A5 levels in T2DM patients, with positive correlations with fasting glucose and HbA1c levels.^[12] These differences may reflect differences in disease stage, treatment exposure, or population-specific genetic differences. A possible biphasic pattern of expression, initially elevated in early metabolic stress but suppressed in more advanced dysfunction, may also explain these divergent results. Longitudinal studies are needed to better characterize these temporal dynamics.

Other studies have linked FAM19A5 to cardiovascular health. Wesołek-Leszczyńska et al associated atherosclerosis-related factors with MetS, highlighting its importance in vascular homeostasis.^[10] Given the cardiovascular risks associated with T2DM, future studies should investigate whether FAM19A5 can serve as an early indicator of cardiovascular complications in this population.

4.3. Potential mechanisms and biological significance

Experimental studies suggest that FAM19A5 may have anti-inflammatory and vasculoprotective effects, including inhibition of neointima formation through the sphingosine-1-phosphate receptor 2 pathway.^[9] In our study, hs-CRP was higher in patients with T2DM, but it was not correlated with FAM19A5 in the pooled sample (ρ = −0.090, P = .354; Table 2). No correlation was found between FAM19A5 and LDL-C (ρ = 0.091, P = .347). These cross-sectional findings do not support a direct anti-inflammatory or lipid-regulatory effect in humans.

Lactate dehydrogenase, platelet count, and WBC were higher in patients with T2DM, consistent with hyperglycemia and low-grade inflammation. However, none was correlated with FAM19A5 in the pooled analysis (Table 2). These nonspecific findings should not be interpreted as evidence of a mechanistic relationship with FAM19A5.

4.4. Strengths, limitations, and future directions

A strength of this study was the inclusion of patients with newly diagnosed, treatment-naïve T2DM, thereby reducing confounding by long disease duration and glucose-lowering treatment. The primary group comparison used a method appropriate to the distribution of FAM19A5, and pooled correlations were reported with adjustment for multiple testing.

This study has several limitations. First, its case-control design and concurrent measurements preclude causal inference; whether lower FAM19A5 is a cause or consequence of T2DM remains unknown. Second, HOMA-IR is less precise than the hyperinsulinemic-euglycemic clamp. Third, the control group was selected for normoglycemia rather than overall metabolic health: 72.2% met the criteria for MetS, the median BMI was 24.9 kg/m² (IQR, 23.7–26.7), and the median HOMA-IR was 2.3. Metabolic dysfunction among controls may have reduced the observed difference in FAM19A5 between groups. More importantly, the absence of a metabolically healthy reference group limits the comparison to newly diagnosed T2DM versus normoglycemia in a metabolically heterogeneous population. Fourth, the ELISA assay has no international calibrator, limiting comparisons of absolute FAM19A5 concentrations across studies. Fifth, the multivariable and subgroup analyses were exploratory, not prespecified, and not internally or externally validated. Finally, the modest sample size (n = 108) limits precision and generalizability. The study was powered for the primary group comparison but not for correlation or subgroup analyses; therefore, nonsignificant findings should not be interpreted as evidence of no association. Larger multicenter longitudinal studies with metabolically healthy controls are needed.

5. Conclusions and future perspectives

In conclusion, FAM19A5 levels were lower in patients with newly diagnosed T2DM than in normoglycemic controls and remained independently associated with diabetes status in an exploratory model that included BMI, HOMA-IR, and hs-CRP. In the pooled sample, FAM19A5 showed weak inverse correlations with glycemic measures but not with BMI, HOMA-IR, lipid measures, or hs-CRP. Subgroup analyses suggested lower FAM19A5 levels in patients with T2DM who had MetS, insulin resistance, or BMI ≥ 25 kg/m², although these findings require confirmation. FAM19A5 alone showed poor discrimination (AUC 0.684; specificity 53.7%), and its addition to the multivariable model did not significantly improve the AUC. These results do not support FAM19A5 as a clinically useful biomarker of early metabolic dysfunction.

The clinical implication for primary care is limited. FAM19A5 cannot currently be recommended for screening or identifying individuals at risk of T2DM; established glycemic and metabolic measures remain the appropriate tools. The findings instead identify FAM19A5 as a candidate adipokine whose regulation may be altered at the time of diabetes diagnosis.

Future studies should clarify the biological pathways involving FAM19A5 and determine whether its modulation affects metabolic outcomes. Such studies should use standardized assays, adequately powered samples covering a broad range of adiposity, and longitudinal designs that can establish temporal relationships.

Acknowledgments

Grammarly was used only for grammar and language checking and not to generate manuscript content.

Author contributions

Conceptualization: Halime Seda Küçükerdem, İsmail Demir.

Funding acquisition: Halime Seda Küçükerdem.

Investigation: Halime Seda Küçükerdem, Giray Bozkaya.

Resources: Halime Seda Küçükerdem, İsmail Demir.

Visualization: Halime Seda Küçükerdem.

Data curation: Giray Bozkaya.

Software: Giray Bozkaya.

Validation: Giray Bozkaya.

Formal analysis: İsmail Demir.

Methodology: İsmail Demir.

Supervision: İsmail Demir.

Writing – original draft: Halime Seda Küçükerdem, İsmail Demir.

Writing – review & editing: Halime Seda Küçükerdem.