Role of fibroblast growth factor-23 in renal disease activity and bone density involvement in patients with lupus nephritis

Abstract

BACKGROUND

Systemic lupus erythematosus-associated lupus nephritis (LN) represents a major contributor to disease-related morbidity. Disturbances in bone metabolism frequently coexist with LN, driven in part by persistent inflammation and impaired renal function. These factors are closely linked to decreased bone mineral density (BMD) and a heightened risk of fractures. Fibroblast growth factor-23 (FGF-23) has recently been recognized as a key regulator of mineral homeostasis and may serve as a marker reflecting both renal disease activity and skeletal involvement. Nevertheless, its precise role in LN activity and its impact on bone health remain inadequately defined.

AIM

To determine the predictive value of FGF-23 in renal disease activity and bone density changes in patients with LN.

METHODS

A cross-sectional hospital-based study included 100 adult patients with LN. Patients were classified into active (n = 50) and inactive LN (n = 50). All patients underwent clinical assessment, laboratory investigations, renal imaging, renal biopsy when indicated, and BMD assessment using dual-energy X-ray absorptiometry. Serum FGF-23 was measured using enzyme-linked immunosorbent assay. Statistical analysis included comparative tests, correlation analysis, logistic regression, and receiver operating characteristic curve analysis.

RESULTS

Patients with active LN had significantly higher FGF-23 levels compared with those with inactive LN (149.44 ± 37.09 pg/mL vs 122.56 ± 36.91 pg/mL; P < 0.001). BMD was significantly lower in patients with active LN (0.42 ± 0.19 g/cm² vs 0.56 ± 0.16 g/cm²; P < 0.001), with low BMD observed in 80% of patients with active LN vs 26% of those with inactive LN. FGF-23 level showed a significant negative correlation with BMD (r = -0.67, P = 0.001). Elevated FGF-23 level independently predicted low BMD (odds ratio = 4.18) and active LN (odds ratio = 2.18). At a cut-off point of ≥ 135 pg/mL, FGF-23 predicted low BMD with 79.5% accuracy.

CONCLUSION

FGF-23 may be a promising biomarker for renal disease activity and BMD in patients with LN. Its elevation may reflect the interplay between disturbances in mineral metabolism, inflammatory processes, and renal involvement. Nevertheless, further large-scale, longitudinal studies are required to validate these observations and clarify the clinical utility of FGF-23 in patients with LN.

Key Words: Bone mineral density; Chronic kidney disease; Fibroblast growth factor-23; Lupus nephritis; Renal disease activity; Systemic lupus erythematosus

Core Tip: Fibroblast growth factor-23 (FGF-23) regulates mineral metabolism and may reflect both renal activity and skeletal involvement. The present study demonstrated that serum FGF-23 levels were significantly associated with renal disease activity and bone mineral density (BMD) in patients with lupus nephritis. Elevated FGF-23 levels were correlated with impaired renal function and reduced BMD. Therefore, FGF-23 may be a valuable biomarker of renal disease activity and the risk of bone loss. Routine assessment of serum FGF-23 in patients with lupus nephritis may aid in risk stratification and disease monitoring. Additionally, those patients should undergo regular evaluation of BMD parameters.

INTRODUCTION

Renal involvement in systemic lupus erythematosus (SLE) is most commonly lupus nephritis (LN), which is one of the most significant systemic involvements. LN represents a prominent source of chronic kidney disease (CKD) progression to end-stage renal disease. It is characterized by immune complex deposition and chronic inflammation within the renal parenchyma, leading to variable degrees of renal impairment[1,2]. A significant percentage of LN patients may have persistent disease activity, renal flare-ups, and long-term consequences that negatively impact survival and quality of life despite advancements in immunosuppressive medication[3].

LN patients are increasingly known to have abnormalities in bone and mineral metabolism[4]. Chronic inflammation, reduced renal function, prolonged glucocorticoid exposure, and immunosuppressive therapies contribute to altered calcium-phosphorus homeostasis and decreased bone mineral density (BMD), raising the risk of fractures and osteoporosis[5]. Identifying biomarkers that reflect both renal disease activity and skeletal involvement is therefore clinically important for early risk stratification and targeted management in this patient population[3,6].

Fibroblast growth factor-23 (FGF-23) is a phosphaturic hormone, and it is synthesized primarily by osteocytes and osteoblasts. FGF-23 is important for regulating phosphate balance, vitamin D synthesis, and bone health. Elevated levels of FGF-23 have been consistently observed in CKD and are closely associated with disease progression, disturbances in mineral metabolism, and unfavorable clinical outcomes[7,8].

Several studies have suggested that FGF-23 may participate in the pathogenesis of renal diseases, including LN. Increased expression of FGF-23 has been observed in glomerular and tubular cells in response to inflammatory stimuli and immune-mediated injury. The growth factor can enhance extracellular matrix production, stimulate mesangial cell proliferation, and promote angiogenesis in the kidney, all of which may contribute to the progression of glomerular damage and renal fibrosis in LN[9].

Furthermore, circulating FGF-23 levels may reflect the extent of renal inflammation and tissue remodeling occurring during active disease. In certain autoimmune and inflammatory diseases, elevated FGF-23 levels have been linked to higher disease activity and severe renal involvement. Therefore, evaluating the role of FGF-23 in patients with LN may provide valuable insights into disease mechanisms. It may help identify potential biomarkers to assess disease activity and predict renal outcomes[10].

Considering these factors, exploring the relationship between FGF-23 levels and LN activity could improve understanding of the molecular pathways involved in kidney injury and help develop new diagnostic or prognostic tools for patients with SLE[11].

However, evidence on the relationship between FGF-23, renal disease activity, and BMD in patients with LN remains limited and inconsistent. Further investigation of this association may help elucidate the pathophysiological interplay between renal inflammation and bone loss in LN[12,13].

The aim of the current study was to assess the relationship between blood FGF-23 levels and both BMD and renal disease activity in LN patients.

MATERIALS AND METHODS

Study design and settings

A cross-sectional design was used for this study. It was carried out in both outpatient clinics and inpatient wards at Assiut University Hospital’s Department of Internal Medicine-Nephrology Division. Patient recruitment and data collection were implemented from March 2025 to December 2025.

Patient selection criteria

Patient eligibility for inclusion required meeting all of the following criteria: (1) Age more than 18 years and less than 60 years; and (2) Diagnosis of SLE based on established classification criteria in the Kidney Disease Improving Global Outcomes guidelines. Presence or absence of LN activity was defined by clinical, laboratory, and histopathological criteria. Patient exclusion was for having end-stage renal disease requiring regular hemodialysis, coexisting autoimmune diseases like rheumatoid arthritis, chronic viral hepatitis B or hepatitis C infection, diabetes mellitus, malignancy, a history of vitamin D supplementation, or any condition known to independently affect bone metabolism.

Sample size calculation and participants’ groups

A total of one hundred adult patients diagnosed with SLE and LN were enrolled in the study. This sample size was considered adequate to detect clinically significant differences in BMD and its predictors in patients with LN, based on the published literature and feasibility considerations. Sample size was calculated with the following assumptions: (1) 80% study power; (2) 5% alpha error; and (3) 95%CI. Based on the LN activity level, patients were assigned to two equal groups: (1) 50 patients with active LN; and (2) 50 patients with inactive LN.

Assessment workups

Every patient had a thorough clinical assessment, which included a targeted medical history of disease duration, medication use, and lupus activity. Additionally, a thorough physical examination was performed, including blood pressure measurement, cardiovascular and respiratory assessments, abdominal examination, assessment of peripheral edema, and bone or neurological involvement. The laboratory investigations included prothrombin time and concentration, international normalized ratio, liver functions, complete blood count, serum creatinine, blood urea nitrogen, and lipid profile.

Dual-energy X-ray absorptiometry scan assessment

BMD was measured by Dual-energy X-ray absorptiometry (DEXA) (OsteoSys Co., Ltd., Korea) to evaluate bone density. The results were expressed in g/cm². Calibration with a machine-specific phantom was performed daily before testing patient samples. Measurements were taken of each subject’s lumbar spine, femoral neck, and total proximal femur. Lumbar spine BMD values were calculated as the average of four measurements for L1-L4.

Human FGF-23

It was measured by enzyme-linked immunosorbent assay with cat. No E0059Hu (Changsheng S Rd, Nanhu Dist, Jiaxing, Zhejiang Province, China).

Abdominal ultrasound

Ultrasonographic evaluation was performed after 8 hours of fasting using a high-resolution ultrasound machine (Aplio; Toshiba Medical Systems Corporation, Tochigi, Japan) with a real-time electronic 3.75 MHz convex-type scanner.

Echocardiography

A transthoracic echocardiographic examination was performed using a high-resolution echocardiography machine with a real-time two-dimensional ultrasound system equipped with a phased-array transducer (2.5-3.5 MHz, Aplio; Toshiba Medical Systems Corporation, Tochigi, Japan).

Renal biopsy

The technique of percutaneous renal biopsy was guided by real-time ultrasound, using an automated spring-loaded biopsy needle. Ethically, informed patient consent was independently obtained for this procedure. The rationale of renal biopsy was to serve as a confirmatory tool of renal disease activity. Biopsy procedures were performed under aseptic conditions. According to the light microscopy findings, immunofluorescence and electron microscopy were further used to examine the processed tissues. Histopathological classification was performed according to the International Society of Nephrology/Renal Pathology Society classification[14].

Study outcomes

The primary outcome was the positive correlation between renal histopathological findings and FGF-23 levels in LN patients. Additionally, the secondary outcome was the positive correlation between low BMD and FGF-23 levels in LN patients.

Ethical considerations

The Ethics Committee of the Faculty of Medicine, Assiut University, Egypt, approved the proposal of this study on March 11, 2025. The institutional review board number is 04-2025-100349.

Statistical analysis

Collected data were analyzed using Statistical Package for the Social Sciences (version 20, IBM, Armonk, NY, United States). The normality of the data was evaluated using the Shapiro-Wilk test. Quantitative data were expressed as mean ± SD. Additionally, quantitative data with normal distributions were compared using the Student t-test, while data with non-normal distributions were compared using the Mann-Whitney U test. χ² test was used to compare nominal data. Correlations between FGF-23 and other variables were tested using Pearson correlation. Logistic regression analysis was used to find predictors of low BMD and active LN. The FGF-23 cut-off for predicting active LN and low BMD was evaluated using the receiver operating characteristic curve. The confidence level was kept at 95%. Hence, a P < 0.05 was considered significant.

RESULTS

Demographic and clinical characteristics of the studied patients

Patients with active LN showed no significant differences compared with those with inactive LN in mean age (P = 0.23), sex (P = 0.18), body mass index (P = 0.78), or number of LN relapses (P = 0.48). However, the duration of LN was significantly longer in patients with active LN (P = 0.02; Table 1).

Table 1 Baseline characteristics of the studied groups, n (%)/mean ± SD.

Variables	Inactive LN group (n = 50)	Active LN group (n = 50)	P value
Age (years)	28.88 ± 5.83	30.56 ± 8.08	0.23
Sex			0.18
Male	17 (34)	12 (24)
Female	33 (66)	38 (76)
Body mass index (kg/m2)	24.57 ± 6.78	23.09 ± 4.48	0.78
Duration of disease (years)	7.68 ± 2.98	9.87 ± 3.45	0.02
Number of relapses	3 (6)	5 (10)	0.48

LN: Lupus nephritis.

The laboratory data of both LN groups are summarized in Table 2. The mean neutrophil/Lymphocyte ratio (NLR) and the platelets/Lymphocytes ratio (PLR) were significantly higher in patients with active LN than in those with inactive LN (P < 0.001). In renal function tests, mean serum creatinine, proteinuria, and estimated glomerular filtration rate were significantly better in patients with active than those with inactive disease (for all, P < 0.001). Additionally, the levels of parathormone (P = 0.02) and complements (C3 and C4; P < 0.001) were significantly different between both groups (Table 2).

Table 2 Laboratory characteristics of the studied groups, n (%)/mean ± SD.

Variables	Inactive LN group (n = 50)	Active LN group (n = 50)	P value
Random blood sugar (mg/dL)	97.42 ± 15.38	108.80 ± 14.31	< 0.001
Complete blood count
Hemoglobin (g/dL)	9.11 ± 1.67	8.86 ± 1.63	0.45
Leucocytes (103/μL)	6.69 ± 2.32	6.98 ± 2.87	0.60
Platelets (103/μL)	233.86 ± 85.29	265.66 ± 106.99	0.97
Neutrophil/Lymphocyte ratio	2.09 ± 0.78	3.11 ± 1.01	< 0.001
Platelets/Lymphocytes ratio	198.11 ± 34.98	256.98 ± 67.89	< 0.001
Kidney function tests
Blood urea nitrogen (mg/dL)	17.90 ± 3.79	19.06 ± 1.90	0.78
Creatinine (μmol/L)	89.36 ± 8.62	232.72 ± 27.50	< 0.001
Proteinuria (mg/day)	1795.30 ± 154.37	3778.14 ± 407.98	< 0.001
Glomerular filtration rate (mL/minute/1.73 m2)	87.41 ± 4.81	57.10 ± 5.26	< 0.001
Urine analysis
Hematuria (HPF)	13.96 ± 3.73	10.07 ± 4.44	0.15
Pyuria (HPF)	10.18 ± 1.56	13.58 ± 2.38	0.93
Albuminuria			0.39
0	6 (1)	4 (8)
+	19 (38)	15 (30)
++	22 (44)	25 (50)
+++	2 (4)	6 (12)
++++	1 (2)	0
Coagulation profile
International normalized ratio	1.04 ± 0.11	1.02 ± 0.17	0.63
Bone parameters
Calcium (mg/dL)	8.07 ± 0.62	7.88 ± 0.66	0.14
Phosphorus (mg/dL)	4.12 ± 1.29	4.71 ± 1.97	0.07
Parathormone (pg/mL)	65.93 ± 17.49	110.24 ± 19.81	0.02
Vitamin D (ng/mL)	37.93 ± 8.52	27.52 ± 7.13	0.22
Alkaline phosphatase (U/L)	95.16 ± 6.33	84.60 ± 9.57	0.36
Complement
C3 (mg/dL)	95.60 ± 12.09	40.18 ± 7.80	< 0.001
C4 (mg/dL)	19.90 ± 3.33	5.55 ± 1.65	< 0.001
Positive antinuclear antibody	50 (100)	50 (100)	-
Positive anti-double-stranded DNA	50 (100)	50 (100)	-

HPF: High-power field; LN: Lupus nephritis.

Radiological data and histopathology among the studied groups

The majority of patients in both groups had normal renal echogenicity, but the frequency of echogenic kidney grades I-III was higher among patients with active LN (48% vs 16%, P = 0.04). One patient with inactive LN had pleural effusion, while 12 (24%) patients with active LN had pleural effusion. However, echocardiography parameters were statistically similar between the two patient groups (Table 3). Renal biopsy confirmed higher disease severity in active LN patients (P < 0.001). Specifically, class IV and V LN were significantly more prevalent among patients with active LN (76%) than among those with inactive LN (8%) (Table 3).

Table 3 Radiological characteristics of the studied groups, n (%)/mean ± SD.

Variables	Inactive LN group (n = 50)	Active LN group (n = 50)	P value
Renal echogenicity			0.04
Normal	42 (84)	26 (52)
Echogenic-I	4 (85)	12 (24)
Echogenic I/II	3 (6)	2 (4)
Echogenic-II	1 (2)	6 (12)
Echogenic-II/III	0	4 (8)
Chest radiograph			< 0.001
Normal findings	49 (98)	38 (76)
Pleural effusion	1 (2)	12 (24)
Echocardiography
Ejection fraction (%)	60.10 ± 10.57	63.24 ± 8.21	0.13
Left ventricular end diastolic diameter (cm)	5.26 ± 1.17	5.13 ± 0.99	0.56
Left ventricular end systolic diameter (cm)	3.50 ± 0.99	3.46 ± 0.73	0.18
Pulmonary artery systolic pressure (mmHg)	34.85 ± 11.93	32.91 ± 6.04	0.09
Left ventricular diastolic dysfunction	3 (8.6)	4 (11.4)	0.40
Renal biopsy LN classes			< 0.001
Class-I	18 (36)	1 (2)
Class-II	28 (56)	9 (18)
Class-III	0	2 (4)
Class-IV	3 (6)	21 (42)
Class-V	1 (2)	17 (34)

LN: Lupus nephritis.

Level of FGF-23 among the studied groups

Patients with active LN had a significantly higher FGF-23 mean level (149.44 ± 37.09 pg/mL vs 122.56 ± 36.91 pg/mL; P < 0.001) compared to those with inactive LN (Figure 1A).

Figure 1 Mean fibroblast growth factor-23 in patients with inactive and active lupus nephritis and in the studied patient groups based on bone mineral density. A: Patients with inactive and active lupus nephritis; B: Studied patient groups based on bone mineral density. BMD: Bone mineral density; LN: Lupus nephritis.

BMD among the studied groups

DEXA scan was significantly lower in patients with active LN (0.42 ± 0.19 g/cm² vs 0.56 ± 0.16 g/cm²) than in those with inactive LN. Additionally, the T score and Z score were significantly different in the two groups (P < 0.001). Furthermore, the majority (74%) of patients with inactive LN had normal BMD, while the majority (80%) of patients with active LN had low BMD (P < 0.001; Table 4).

Table 4 Bone mineral density among the studied groups, n (%)/mean ± SD.

Variables	Inactive LN group (n = 50)	Active LN group (n = 50)	P value
Dual-energy X-ray absorptiometry scan (g/cm2)	0.56 ± 0.16	0.42 ± 0.19	< 0.001
T score	- 0.36 ± 0.15	-1.53 ± 0.18	< 0.001
Z score	- 0.28 ± 0.14	-1.49 ± 0.22	< 0.001
Class of BMD			< 0.001
Normal BMD	37 (74)	10 (20)
Low BMD	13 (26)	40 (80)

BMD: Bone mineral density; LN: Lupus nephritis.

Characteristics of patient groups based on BMD

Low BMD patients had significantly lower body mass index and longer disease duration (P < 0.001). Frothy urine was far more common in those with low BMD (P = 0.002), lower limb edema (P = 0.002), and high LN classes in renal biopsy (P < 0.001; Table 5).

Table 5 Characteristics of the studied patients based on bone mineral density status, n (%)/mean ± SD.

Variables	Normal BMD (n = 47)	Low BMD (n = 53)	P value
Age (years)	28.72 ± 6.58	30.60 ± 7.40	0.18
Sex			0.47
Male	13 (27.7)	16 (30.2)
Female	34 (72.3)	37 (69.8)
Body mass index (kg/m2)	25.91 ± 6.11	21.09 ± 6.89	< 0.001
Duration of disease (years)	5.67 ± 2.67	10.46 ± 3.11	< 0.001
Frothy urine	5 (10.6)	20 (37.7)	0.002
Oliguria/anuria	4 (8.5)	10 (18.9)	0.11
Lower limb edema	6 (12.8)	20 (37.7)	0.002
New-onset hypertension	2 (4.3)	7 (13.2)	0.11
Pulmonary congestion	1 (2.1)	5 (9.4)	0.13
Raised renal chemistry	1 (2.1)	4 (7.5)	0.22
Anasarca	1 (2.1)	1 (1.9)	0.72
Volume overload	1 (2.1)	0 (0.0)	0.47
Renal biopsy lupus nephritis classes			< 0.001
Class-I	16 (34)	3 (5.7)
Class-II	23 (48.9)	14 (26.4)
Class-III	0	2 (3.8)
Class-IV	3 (6.4)	21 (39.6)
Class-V	5 (10.6)	13 (24.5)

BMD: Bone mineral density.

Comparative results of patient groups based on changes in BMD

Patients with low BMD have significantly higher NLR, PLR, serum creatinine (P < 0.001), and proteinuria (P = 0.01) compared to those with normal BMD. At the same time, those patients with low BMD had significantly lower vitamin D3, C3, and C4 levels (Table 6). Patients with low BMD had significantly higher FGF-23 (150.76 ± 32.45 pg/mL vs 119.36 ± 39.82 pg/mL, P < 0.001) compared to those with normal BMD (Figure 1B).

Table 6 Laboratory characteristics of the studied patients based on bone mineral density status, n (%)/mean ± SD.

Variables	Normal BMD (n = 47)	Low BMD (n = 53)	P value
Random blood sugar (mg/dL)	98.26 ± 15.06	107.42 ± 15.41	< 0.001
Complete blood count
Hemoglobin (g/dL)	9.25 ± 1.84	8.75 ± 1.41	0.13
Leucocytes (103/μl)	7.14 ± 2.40	6.54 ± 2.76	0.25
Platelets (103/μL)	293.74 ± 87.80	248.49 ± 104.66	0.02
Neutrophil/Lymphocyte ratio	1.22 ± 0.38	2.67 ± 0.56	< 0.001
Platelets/Lymphocytes ratio	156.89 ± 34.90	245.09 ± 56.90	< 0.001
Kidney function tests
Blood urea nitrogen (mg/dL)	19.14 ± 4.03	17.89 ± 1.76	0.78
Creatinine (μmol/L)	100.37 ± 13.97	214.60 ± 25.65	< 0.001
Proteinuria (mg/day)	2158.98 ± 276	3343.40 ± 362.98	0.01
Glomerular filtration rate (mL/minute/1.73 m2)	81.92 ± 32.64	63.68 ± 41.67	0.01
Urine analysis
Hematuria (HPF)	13.26 ± 3.64	9.23 ± 1.98	0.31
Pyuria (HPF)	10.11 ± 3.47	13.46 ± 2.31	0.24
Albuminuria			0.62
0	6 (12.8)	4 (7.5)
+	17 (36.2)	17 (32.1)
++	20 (42.6)	27 (50.9)
+++	3 (6.4)	5 (9.4)
++++	1 (2.1)	0 (0.0)
Coagulation profile
Prothrombin time (seconds)	11.88 ± 1.53	12.21 ± 176	0.43
Prothrombin concentration (%)	93.15 ± 13.93	94.11 ± 17.87	0.81
International normalized ratio	1.03 ± 0.17	1.03 ± 0.09	0.97
Bone parameters
Calcium (mg/dL)	8.01 ± 0.64	7.94 ± 0.65	0.55
Phosphorus (mg/dL)	4.07 ± 1.19	4.23 ± 1.98	0.56
Parathormone (pg/mL)	72.64 ± 33.56	101.22 ± 18.61	0.16
Vitamin D (ng/mL)	39.90 ± 9.05	22.36 ± 4.84	0.01
Alkaline phosphatase (U/L)	94.47 ± 6.45	85.81 ± 9.26	0.45
Complement
C3 (mg/dL)	100.98 ± 13.99	50.56 ± 12.98	< 0.001
C4 (mg/dL)	23.89 ± 6.78	9.05 ± 1.90	< 0.001

BMD: Bone mineral density; HPF: High-power field.

Correlation of FGF-23 with other variables

FGF-23 has a significant negative correlation with DEXA scan (Table 7).

Table 7 Correlation of fibroblast growth factor-23 with the other variables, n (%).

Variables	Fibroblast growth factor-23
Age (years)	-0.01 (0.90)
Body mass index (kg/m2)	0.03 (0.75)
Duration	-0.14 (0.20)
Random blood sugar (mg/dL)	-0.10 (0.34)
Hemoglobin (g/dL)	-0.03 (0.79)
Leucocytes (103/μL)	0.13 (0.13)
Platelets (103/μL)	0.12 (0.82)
Neutrophil/Lymphocyte ratio	-0.15 (0.17)
Platelets/Lymphocytes ratio	-0.19 (0.20)
Blood urea nitrogen (mg/dL)	0.10 (0.34)
Creatinine (μmol/L)	0.21 (0.06)
Proteinuria (mg/day)	0.01 (0.26)
Glomerular filtration rate (mL/minute/1.73 m2)	0.09 (0.51)
Urine analysis
Hematuria (HPF)	0.06 (0.53)
Pyuria (HPF)	-0.04 (0.77)
Albuminuria	-0.18 (0.48)
Coagulation profile
Prothrombin time (seconds)	-0.02 (0.86)
Prothrombin concentration (%)	-0.15 (0.17)
International randomized ratio	0.10 (0.34)
Bone parameters
Calcium (mg/dL)	0.14 (0.26)
Phosphorus (mg/dL)	0.12 (0.51)
Parathormone (pg/mL)	-0.11 (0.08)
Vitamin D (ng/mL)	0.12 (0.53)
Alkaline phosphatase (U/L)	-0.19 (0.17)
Complement
C3 (mg/dL)	-0.14 (0.92)
C4 (mg/dL)	-0.08 (0.48)
Echocardiography
Ejection fraction (%)	0.10 (0.38)
Left ventricular end diastolic diameter (cm)	0.03 (0.76)
Left ventricular end systolic diameter (cm)	-0.18 (0.10)
Pulmonary artery systolic pressure (mmHg)	-0.05 (0.62)
Dual-energy X-ray absorptiometry scan (g/cm2)	-0.67 (0.001)

HPF: High-power field.

Regression analysis for the prediction of low BMD in patients with LN

According to the present study, predictors for low BMD in LN patients included active LN [odds ratio (OR) = 6.95)], advanced LN class (OR = 3.69), and elevated FGF-23 (OR = 4.18) (Table 8). At a cut-off > 135 pg/mL, FGF-23 has 79.5% overall accuracy in predicting low BMD in patients with LN, with an area under the curve of 0.83 (Table 9, Figure 2A).

Figure 2 Receiver operating characteristics curve for predictive value of fibroblast growth factor-23 in the prediction of low bone mineral density in patients with lupus nephritis and active lupus nephritis. A: The prediction of low bone mineral density in patients with lupus nephritis; B: The prediction of active lupus nephritis.

Table 8 Regression analysis for prediction of low bone mineral density in patients with lupus nephritis.

Predictor	Odds ratio	95%CI	P value
Age	0.47	0.33-1.68	0.56
Sex	1.12	0.78-2.22	0.08
Low body mass index	1.42	0.18-1.88	0.68
Disease duration (years)	1.36	0.67-1.69	0.77
Active LN	6.95	1.89-14.21	< 0.001
Serum creatinine (μmol/L)	1.04	0.56-1.08	0.66
Glomerular filtration rate (per 10 mL/minute decrease)	0.69	0.12-2.09	0.90
Proteinuria (per 1 g/day)	1.88	0.59-3.10	0.58
Neutrophil/Lymphocyte ratio	0.41	0.36-4.92	0.71
Platelets/Lymphocytes ratio	0.68	0.71-1.04	0.89
Low C3	0.72	0.38-7.95	0.18
Low C4	0.21	0.17-2.42	0.46
Advanced LN class (III-V)	3.69	2.11-10.34	< 0.001
Fibroblast growth factor-23 (per 10 pg/mL increase)	4.18	1.49-9.90	< 0.001

LN: Lupus nephritis.

Table 9 Receiver operating characteristic curve analysis of fibroblast growth factor-23 for the prediction of low bone mineral density.

Parameter	Value
Area under the curve	0.83
95%CI	0.75-0.91
P value	< 0.001
Optimal cutoff value, pg/mL	≥ 135
Sensitivity (%)	81
Specificity (%)	78
Positive predictive value (%)	80
Negative predictive value (%)	79
Accuracy (%)	79.5

Regression analysis for the prediction of active LN

Based on the current study, predictors of active LN were NLR (OR = 1.47), PLR (OR = 1.69), and elevated FGF-23 (OR = 2.18) (Table 10). At a cut-off > 121.1 pg/mL, FGF-23 has 80% overall accuracy in the prediction of active LN, with an area under the curve of 0.86 (Table 11, Figure 2B).

Table 10 Regression analysis for the prediction of active lupus nephritis.

Predictor	Odds ratio	95%CI	P value
Age	1.13	0.34-1.67	0.44
Sex	1.01	0.88-2.01	0.12
Low body mass index	0.67	0.34-1.34	0.40
Disease duration (years)	0.98	0.22-1.45	0.60
Serum creatinine (μmol/L)	0.67	0.11-1.68	0.76
Glomerular filtration rate (per 10 mL/minute decrease)	1.30	0.59-2.11	0.21
Proteinuria (per 1 g/day)	1.22	0.70-2.44	0.89
Neutrophil/Lymphocyte ratio	0.78	0.54-2.18	0.46
Platelets/Lymphocytes ratio	1.47	1.11-1.89	0.01
Low C3	1.69	1.49-1.19	0.01
Low C4	1.10	0.70-2.32	0.09
Fibroblast growth factor-23 (per 10 pg/mL increase)	2.18	1.22-4.45	< 0.001

Table 11 Receiver operating characteristic curve analysis of fibroblast growth factor-23 for the prediction of active lupus nephritis.

Parameter	Value
Area under the curve	0.86
95%CI	0.78-0.94
P value	< 0.001
Optimal cutoff value, pg/mL	≥ 121.10
Sensitivity (%)	79
Specificity (%)	81
Positive predictive value (%)	80.6
Negative predictive value (%)	79.4
Accuracy (%)	80.0
Disease frequency (%)	50

DISCUSSION

FGF-23 is a phosphaturic hormone primarily produced by osteocytes and osteoblasts. It plays a crucial role in regulating phosphate homeostasis and vitamin D metabolism. In the context of LN, immune-mediated renal injury may disrupt normal phosphate handling and stimulate increased FGF-23 production. Additionally, inflammatory cytokines commonly elevated during disease flares may further enhance FGF-23 secretion, linking mineral metabolism disturbances with the inflammatory activity of LN[15].

The current study showed that serum FGF-23 levels were significantly elevated in patients with active LN. Higher concentrations of FGF-2 were also associated with reduced BMD, suggesting a potential interplay among renal inflammation, disordered mineral metabolism, and skeletal involvement in LN. These observations indicate that FGF-23 may serve as a biomarker reflecting both renal disease severity and bone health in this patient population, in line with previous reports[16].

In the present study, serum FGF-23 levels were significantly elevated in patients with active LN compared with those in an inactive state. This finding further supports accumulating evidence linking FGF-23 to renal inflammatory processes and progressive kidney dysfunction. Prior studies have shown that FGF-23 levels increase early in CKD and correlate with disease severity, even before detectable alterations in serum phosphate levels occur[10,11].

The observed increase in serum FGF-23 levels in active LN patients suggests that this biomarker may have potential clinical value in reflecting disease activity and renal involvement. High levels of FGF-23 have been reported with worsening kidney function, high grades of proteinuria, and greater degrees of renal damage. Therefore, monitoring FGF-23 levels may help identify patients with more aggressive disease and contribute to early detection of renal deterioration. Furthermore, FGF-23 might serve as a non-invasive biomarker for assessing disease progression and response to therapy in LN patients, complementing traditional laboratory markers such as serum creatinine, complement levels, and anti-double-stranded DNA antibodies[10].

Experimental and clinical evidence indicates that inflammatory cytokines, which are commonly elevated in active LN, can directly stimulate FGF-23 production. In addition, reduced renal clearance and potential tubular resistance to FGF-23 may further contribute to its accumulation during active renal disease. Similar observations were reported by Gutiérrez et al[17], who demonstrated significantly higher FGF-23 levels in patients with inflammatory kidney diseases compared with controls, independent of estimated glomerular filtration rate.

In the present study, elevated FGF-23 levels showed significant correlations with impaired renal function indices, including increased serum creatinine and reduced estimated glomerular filtration rate. These findings are consistent with previous reports indicating that FGF-23 rises progressively with declining renal function and reflects both reduced phosphate excretion and underlying renal parenchymal injury[18,19].

In the context of LN, proteinuria and active glomerular inflammation may further disrupt phosphate handling, thereby amplifying FGF-23 secretion. This highlights the role of FGF-23 not merely as a marker of mineral metabolism but also as an indicator of renal disease burden[20].

A previous study found an association between elevated serum FGF-23 levels and tissue damage in patients with SLE, particularly the musculoskeletal complications and proteinuria. This would imply that FGF-23 plays a role in the pathophysiology of the disease. The role is either causal or a consequence of the immunoinflammatory alterations in SLE. Accordingly, FGF-23 might be regarded as a biomarker of musculoskeletal and renal damage in SLE patients[21].

This was consistent with research suggesting that an increase in FGF-23 is among the earliest detectable indicators of CKD[22]. Of note, the increase in FGF-23 levels in CKD occurs even before changes in parathormone, serum calcium, or phosphorus levels[23]. In renal proximal tubular cells, FGF-23 downregulates the luminal membrane sodium phosphate cotransporter. This mechanism is mediated by the binding of FGF-23 to its receptor and klotho as a coreceptor. Suppressed cotransporters in the proximal tubule restrict phosphate reabsorption and boost urine phosphate excretion. Additionally, FGF-23 suppresses 1-alpha-hydroxylase activity in the proximal renal tubule, limiting renal synthesis of calcitriol[21].

An important finding of the current study is the significant association between elevated FGF-23 levels and reduced BMD. Patients with higher FGF-23 concentrations exhibited lower BMD values, indicating an increased risk of osteopenia and osteoporosis. This relationship may be mediated by the inhibitory effect of FGF-23 on active vitamin D synthesis, which in turn reduces calcium absorption and promotes secondary hyperparathyroidism[20,24].

Previous studies in CKD and autoimmune disorders have similarly demonstrated inverse associations between FGF-23 levels and bone density[16]. In patients with LN, additional factors – including prolonged corticosteroid use, persistent inflammation, and vitamin D deficiency – may further exacerbate the adverse skeletal effects associated with elevated FGF-23 levels[25].

These findings carry important clinical implications. Elevated FGF-23 may serve as a useful biomarker for identifying active LN in patients at heightened risk of bone loss. Early detection of disturbances in mineral metabolism could enable timely interventions, such as optimizing vitamin D status, reducing corticosteroid exposure where feasible, and implementing targeted strategies for bone health monitoring[25-27]. Furthermore, monitoring FGF-23 levels may provide complementary insight into disease activity beyond conventional renal biomarkers, potentially enhancing risk stratification and improving long-term clinical outcomes[27,28].

The connection between high FGF-23 levels and the inflammatory biomarkers in LN patients has been previously described. Certain mechanisms for this association are incompletely understood, but numerous possible explanations exist. The notion is that inflammation could increase FGF-23 synthesis by the osteocytes. In support of this, it has been proven that inflammation-induced bone resorption can enhance the FGF-23 production. Additionally, the ensuing bone resorption leads to a local increase in phosphate and calcium, which may promote FGF-23 synthesis[29].

Among the limitations of the current study are its cross-sectional design, which precludes causal inference, and the relatively limited sample size. Longitudinal studies are warranted to evaluate whether changes in FGF-23 levels predict renal flares, LN progression, or future bone loss. Further studies are warranted to determine whether targeted modulation of FGF-23 may lead to improvements in renal and skeletal outcomes in this patient population.

CONCLUSION

The present study demonstrated that serum FGF-23 levels were significantly associated with renal disease activity and BMD in patients with LN. Higher FGF-23 levels were detected in patients with active renal disease and were correlated with impaired renal function and lower BMD. These findings suggest that FGF-23 reflects the interaction between renal inflammation, disturbances in mineral metabolism, and skeletal involvement in LN. Accordingly, FGF-23 seems to be a useful biomarker for assessing disease activity and identifying patients at increased risk of bone loss. Regular evaluation of serum FGF-23, particularly in patients with active disease or declining renal function, may aid in early detection, risk stratification, disease monitoring, and timely management of BMD.