**3. Results**

#### *3.1. Baseline Characteristics of the Study Subjects Stratified by BMI*

The mean age of the subjects was 41.2 ± 14.7 years, and the mean BMI was 24.1 ± 4.1 kg/m2. Among 537 patients, 32 patients (5.9%) were underweight, 312 patients (57.8%) were normal weight or overweight, and 193 patients (35.7%) were obese. The baseline clinical characteristics of the study population were compared between the three groups according to BMI classification (Table 1). The obese group was older and had a higher prevalence of hypertension than the underweight and normal weight/overweight groups. SBP and DBP and blood levels of hemoglobin, fasting glucose, liver enzymes, uric acid, total cholesterol, triglycerides, LDL-C, 24-h proteinuria, C3, and C4 were higher in the obese group compared to the other groups, while the eGFR was lower. Figure S2 shows the distribution of BMI according to age and sex categories. For all patients, the BMI distribution peak was shifted to the left. Patients with BMI ≥ 25 kg/m<sup>2</sup> accounted for 35.9% of the total subjects (40.4% in males and 31.8% in females). Patients who had lower BMI tended to be younger, and the distribution of BMI was similar between males and females.


**Table 1.** Baseline clinical variables of the three BMI groups at the time of renal biopsy.

\* ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; C3, complement 3; C4, complement 4; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL-C, high density lipoprotein-cholesterol; hs-CRP, high sensitivity-C-reactive protein; IgA, immunoglobulin A LDL-C, low density lipoprotein-cholesterol; RBCs, red blood cells; SBP, systolic blood pressure. \* *p* < 0.05 vs. BMI < 18.5 kg/m<sup>2</sup> and † *p* < 0.05 vs. BMI 18.5–24.9 kg/m<sup>2</sup> by one-way ANOVA with Scheffe's post-hoc analysis.

#### *3.2. Histopathologic Findings of the Study Subjects Stratified by BMI*

We compared histopathologic features between the three BMI groups in Tables 2 and 3 and Figure 1. Total number of glomeruli and the mean mesangial C3 deposition score were lower (*p* = 0.003 and *p* < 0.001, respectively), and the mean MME (*p* = 0.042), IF (*p* = 0.046) and TA (*p* = 0.033) score was higher in the obese group compared to the other groups (Table 2). On light microscopy, the glomeruli, mesangium, tubules, interstitium, vessels, and mesangial IgA deposition score did not differ between the three groups. There was an increasing trend of high grade MME (*p* = 0.007), IF (*p* = 0.03), and TA (*p* = 0.039) as the BMI increased. The distribution of the C3 deposition severity was different among the three groups (*p* < 0.001, Figure 1). The distributions for WHO and Oxford classifications of IgAN were not significantly different between the three BMI groups (Table 3).


**Table 2.** Histopathological findings of the three BMI groups.

\* BMI, body mass index; C3, complement 3; C4, complement 4; IgA, immunoglobulin A. \* *p*<0.05 vs. BMI<18.5 kg/m<sup>2</sup> and † *p* < 0.05 vs. BMI 18.5–24.9 kg/m<sup>2</sup> by post-hoc analysis of Kruskal-Wallis test.


**Table 3.** The distribution of histological classifications according to the three BMI groups.

BMI, body mass index; WHO, World Health Organization. Oxford classification: M; mesangial hypercellularity, E; endocapillary proliferation, S; segmental sclerosis, T tubular atrophy/interstitial fibrosis. Statistical analysis was performed by Kruskal-Wallis test in WHO classifications and using the chi-squared test in Oxford classifications.

**Figure 1.** The distribution of histopathologic scores according to the three BMI groups. (**a**) Mesangial matrix expansion (MME). (**b**) Mesangial cell proliferation (MCP). (**c**) Interstitial fibrosis (IF). (**d**) Tubular atrophy (TA). (**e**) IgA mesangial deposition. (**f**) C3 mesangial deposition.

#### *3.3. Association between BMI and Histopathologic Parameters*

Univariable linear regression analysis showed that scores for MME, endocapillary proliferation, IF, and TA were positively correlated with BMI, while the number of total glomeruli and the mesangial C3 deposition score were negatively correlated with BMI (Table 4). Multivariable linear regression analysis showed that the MME and mesangial IgA deposition scores were positively associated with BMI (*p* = 0.028; adjusted R<sup>2</sup> = 0.291 and *p* = 0.041; adjusted R<sup>2</sup> = 0.291, respectively), while total number of glomeruli was negatively associated with BMI (*p* = 0.029; adjusted R<sup>2</sup> = 0.286) after adjusting for clinical parameters including age, SBP, hemoglobin, glucose, albumin, AST, ALT, uric acid, total cholesterol, eGFR, 24-h proteinuria, and serum C3, C4, and IgA levels.


**Table 4.** Linear regression analysis for BMI and the histopathologic parameters.

BMI, body mass index, C3, complement 3; C4d, cleavage product of complement 4; IgA, immunoglobulin A. Multivariable analysis was adjusted for each histologic parameter and clinical parameters, including age, systolic blood pressure, hemoglobin, glucose, albumin, AST, ALT, uric acid, total cholesterol, eGFR, 24-h proteinuria, and serum C3, C4 and IgA levels.

Table 5 shows the OR and 95% CI for high MME, IF, TA, and positive mesangial deposition of C3 and IgA for the three BMI groups. Considering the normal weight or overweight group as a reference, the obese group showed higher risks for high MME and TA, and lower risk for mesangial C3 deposition. After adjusting for age, sex, the presence of hypertension and diabetes mellitus, and SBP (Model 1), the obese group exhibited higher risk for severe MME (OR = 2.066, 95% CI 1.227–3.478, *p* = 0.006) and lower risk for mesangial C3 deposition (OR = 0.544, 95% CI 0.365–0.810, *p* = 0.003). After adjusting for model 1 with total glomerular number, eGFR, 24-h proteinuria, and blood levels of hemoglobin, glucose, albumin, AST, ALT, uric acid, total cholesterol, C3, C4, and IgA (Model 2), the obese group still had a significantly higher risk for severe MME (OR = 2.060, 95% CI 1.120–3.788, *p* = 0.020), while the underweight group had a significantly lower risk for severe MME (OR = 0.369, 95% CI 0.150–0.904, *p* = 0.029) and mesangial IgA deposition (OR = 0.208, 95% CI 0.049–0.889, *p* = 0.034).


**Table 5.** Logistic regression analysis for BMI groups and histopathologic parameters.

BMI, body mass index; C3, complement; IgA, immunoglobulin A. Model 1 was adjusted for age, sex, the presence of hypertension, the presence of diabetes mellitus and systolic blood pressure. Model 2 was adjusted for model 1 + total glomerular number, eGFR, 24-h proteinuria, and blood levels of hemoglobin, glucose, albumin, aspartate aminotransferase, alanine aminotransferase, uric acid, total cholesterol, and serum C3, C4, and serum IgA levels.

To evaluate whether the association between BMI and MME in IgAN is an incidental finding in obese patients, we analyzed subjects with tubulointerstitial disease without glomerular injury (acute tubular necrosis, interstitial nephritis, and pyelonephritis) during the same period in our kidney biopsy cohort (n = 80). There was no significant association between BMI groups and MME in subjects with tubulointerstitial diseases (Table S1 (Supplementary Materials)).

#### *3.4. Association of BMI with Clinical Variables, Renal Function, and Proteinuria*

There was a positive relationship between BMI and 24 h proteinuria, total cholesterol, triglycerides, and serum C3 and C4 levels and a negative relationship between BMI and eGFR (Figure 2). The association of BMI with renal function and proteinuria was evaluated by adjusting for clinical and histopathological variables in linear regression analysis (Table 6). BMI was negatively correlated with eGFR in univariable analysis; however, this relationship was not significant on multivariable analysis. Factors significantly associated with eGFR were age, SBP, levels of serum albumin and HDL-C, total number of glomeruli, and the interstitial fibrosis score (adjusted R<sup>2</sup> = 0.460). BMI showed a positive linear association with 24-h proteinuria on univariable analysis, but the association was not significant on multivariable analysis. eGFR and serum levels of albumin and C3 were significantly associated with 24-h proteinuria (adjusted R<sup>2</sup> = 0.260).



the univariable linear regression (*p* < 0.1). Dashes indicate that the variable did not enter the multivariable linear regression model.

**Figure 2.** The correlation graphs between BMI and clinical variables. (**a**) CKD-EPI eGFR. (**b**) 24 h proteinuria. (**c**) Total cholesterol level. (**d**) Triglyceride level. (**e**) Serum C3 level. (**f**) Serum C4 level.

#### *3.5. Treatment Patterns of the Study Subjects Stratified by BMI*

At the time of kidney biopsy, there was no significant difference in the use of renin-angiotensinaldosterone system (RAAS) blockers among the three groups (*p* = 0.252); BMI < 18.5 kg/m<sup>2</sup> (n = 13, 40.6%) vs. 18.5–24.9 kg/m<sup>2</sup> (n = 87, 27.9%) vs. ≥ 25 kg/m<sup>2</sup> (n = 51, 26.4%). Several treatment strategies were chosen within 6 months after kidney biopsy and these were compared between the three groups. More patients in the obese group were treated with RAAS blockers, calcium channel blockers (CCBs), β-blockers, thiazide, statins, and corticosteroids compared to the underweight and normal/overweight groups (Table 7).

**Table 7.** Distribution of treatment after renal biopsy according to the three BMI groups.


BMI, body mass index; CCB, calcium channer blocker; RAAS, renin-angiotensin-aldosterone system.
