Association of Obesity with Cardiovascular Risk Factors and Kidney Disease Outcomes in Primary Proteinuric Glomerulopathies

Abstract

Background/Aims:

Obesity is a known risk factor for cardiovascular disease and contributes to the development and progression of kidney disease. However, the specific influence of obesity on outcomes in primary glomerular disease has not been well characterized.

Methods:

In this prospective cohort study, data were from 541 participants enrolled in the Nephrotic Syndrome Study Network (NEPTUNE), between 2010 and 2019, at 23 sites across North America. Blood pressure, lipids, and kidney disease outcomes including complete proteinuria remission, kidney failure, and chronic kidney disease progression were evaluated. Data were analyzed using linear and logistic regression with generalized estimating equations and time-varying Cox regression with Kaplan-Meier plots.

Results:

The prevalence of obesity at baseline was 43.3% (N=156) in adults and 37.6% (N=68) in children. In adults, obesity was longitudinally associated with higher systolic BP (β=6.49, 95% CI: 2.41,10.56, p=0.002), dyslipidemia (OR=1.74, 95% CI: 1.30,2.32, p<0.001), triglycerides (β=41.92, 95% CI: 17.12,66.71, p=0.001), and lower HDL (β=−6.92, 95% CI: −9.32,−4.51, p<0.001). In children, obesity over time was associated with higher systolic BP index (β=0.04, 95% CI: 0.02,0.06, p<0.001) and hypertension (OR=1.43, 95% CI: 1.04,1.98, p=0.03). In both adults and children, obesity was associated with a significantly lower hazard of achieving complete remission of proteinuria (adult HR=0.80, 95% CI: 0.69,0.88, p<0.001, pediatric HR=0.72, 95% CI: 0.61,0.84, p<0.001).

Conclusion:

Obesity was associated with higher cardiovascular risk and less proteinuria remission from nephrotic syndrome in adults and children with proteinuric glomerulopathies. Weight-loss strategies may forestall cardiovascular disease and progressive kidney function decline in this high-risk patient group.

Introduction

The prevalence of obesity in the United States (US) has risen to 39.8% among adults and 18.5% among children [1]. By 2030, more than 50% of the overall US population will be at risk for developing obesity-related health complications [2]. Obesity has been associated with an increased propensity for developing chronic medical conditions, including type 2 diabetes, cardiovascular disease (CVD), depression, and certain cancers. Children with obesity are also at risk for the same chronic conditions, which may develop earlier than among those without childhood onset obesity [3]. Importantly, obesity is a known risk factor for chronic kidney disease (CKD) and has been shown to accelerate CKD progression [4] and graft loss after kidney transplantation [5] [6]. In addition, obesity can result in a secondary form of focal segmental glomerulosclerosis (FSGS), termed obesity-related glomerulopathy [7]. As such, previous studies have shown that the incidence of glomerulopathy has been rising in conjunction with the global obesity epidemic [8].

This association of obesity with glomerulopathy is thought to be related to glomerular hyperfiltration and elevated intraglomerular pressure from the increased metabolic demand present among individuals with obesity [4]. Pathophysiologically, increased fat mass is thought to cause mesangial expansion [9] and increase renal metabolic demand [10], promoting glomerular hyperfiltration [11], glomerular hypertrophy [12], decreased podocyte density [13], and increased filtration fraction [14]. Moreover, as podocytes cannot proliferate independently, a decrease in podocyte density disproportionately increases mechanical strain on these cells, leading to glomerular scarring [14]. These processes may collectively contribute to the development and progression of CKD.

Primary proteinuric glomerulopathies such as minimal change disease (MCD), FSGS and membranous nephropathy (MN) often manifest as nephrotic syndrome. Treatment focuses on decreasing proteinuria and inducing remission as a means to decrease symptomology and prevent or slow CKD progression; however, typical clinical courses may include repeating cycles of remission and relapse of nephrotic syndrome despite treatment. The influence of obesity on kidney disease outcomes in primary proteinuric glomerulopathies has not been well characterized. One study using a Japanese registry of kidney biopsies found that greater body mass index (BMI) was associated with proteinuria in MCD and MN, but not in FSGS [15]. Another single center study found that BMI was not related to progression of kidney disease [16]. Further, although the prevalence of obesity in patients with glomerular disease has been reported in a handful of studies [17–19], the association between obesity and cardiovascular risk profile has not been previously described.

Herein, the purpose of this study was to further examine the association between obesity, cardiovascular risk factors, and kidney disease outcomes among adults and children with primary proteinuric glomerulopathies who were enrolled in a large multi-center cohort, the Nephrotic Syndrome Study Network (NEPTUNE). We sought to determine the prevalence of obesity in adults and children with biopsy proven MCD, FSGS and MN, and to investigate the association of obesity with blood pressure (BP), lipids, CKD progression, kidney failure, nephrotic syndrome remission and pathology findings among adults and children with primary proteinuric glomerular disease.

Materials and Methods

Nephrotic Syndrome Study Network (NEPTUNE)

NEPTUNE is a National Institutes of Health-sponsored, multi-center observational cohort study of adults and children with glomerular diseases associated with nephrotic syndrome. The design of the NEPTUNE study has been previously described in detail [20]. Participants were enrolled in two waves, NEPTUNE 1 and NEPTUNE 2. For NEPTUNE 1, participants with proteinuria ≥500 mg/day on a 24-hour urine sample or urine protein/creatinine ratio (UPC) ≥0.5 g/g on a spot urine specimen were enrolled at the time of kidney biopsy at one of 23 sites across North America. To refine the cohort to primary glomerular disorders, NEPTUNE 2 used a UPC >1.5 g/g as the inclusion criteria. Individuals presenting with kidney manifestations of systemic disease, history of solid organ transplant, or life expectancy <6 months were excluded from NEPTUNE. Based on histological determination by core pathologists, participants were divided into disease cohorts of MCD, FSGS, MN, or other glomerulopathy. At each study visit, demographic data, clinical characteristics, vital signs, physical assessment and blood/urine biosamples of participants were collected. Study visits occurred every 4 months during the initial year, followed by every 6 months for a total of 5 years. Participants enrolled between July 1, 2010 and June 12, 2019 were included in this analysis. The Institutional Review Board (IRB) at each participating site approved the study protocol, and informed consent/assent was obtained from each participant.

Anthropometric Measurements

Height was measured to the nearest 0.1 cm with a stadiometer and weight to the nearest 0.1 kg with a digital scale. Participants’ weight was classified at each study visit as obese or non-obese based on reference data for BMI in adults ≥20 years or BMI percentile in children (<20 years). The criteria used to define obesity were BMI ≥30 kg/m² for adults and BMI ≥95^th percentile for children [1].

Cardiovascular Measures

Casual BP measurements were obtained in triplicate at each study visit using a validated oscillometric device. BP was measured in the right arm with the participant in a seated position after five minutes of rest. The average of the last two readings was used in analyses. Hypertensive BP was defined as systolic BP ≥130 mmHg or diastolic BP ≥80 mmHg for adults and children ≥13 years of age according to clinical practice guidelines [21]. Systolic or diastolic BP ≥95^th percentile was considered to be in the hypertensive BP range in children <13 years [21]. In order to compare BP across different age groups, BP index (mean BP/hypertensive BP) was calculated in children and BP index ≥1 was indicative of hypertensive BP. Participants were classified as “Hypertensive BP Status” (HTN) at each study visit if their average BP was in the hypertensive range or if a clinical diagnosis of hypertension was documented in their medical record.

Total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol and triglycerides were measured centrally using non-fasting blood samples. Dyslipidemia at each study visit was defined by any of the following: HDL <40 mg/dL; non-HDL ≥145 mg/dl (<18 years), ≥160 mg/dl (≥18 years); or triglycerides ≥100 mg/dL (0–9 years), ≥130 mg/dL (10–17 years), or ≥150 mg/dL (≥18 years) [22].

Kidney Disease Outcomes

Kidney disease outcomes were pre-specified in the NEPTUNE protocol and have been previously reported [20]. Kidney failure was defined as either two consecutive study visits with estimated glomerular filtration rate (eGFR) values <15 ml/min/1.73m² or a diagnosis of End Stage Renal Disease (ESRD; dialysis or transplant). CKD progression was defined as having both an eGFR decline by 40% and an eGFR <90 ml/min/1.73m², or developing ESRD. Complete remission ever was defined as a urine protein to creatinine ratio (UPC) <0.3 g/g at any time point after diagnosis. eGFR was calculated using the CKD-EPI formula for adults and the modified Schwartz formula for children [23, 24], except for those between 18–26 years old where we averaged the CKD-Epi and Schwartz formulas [25].

Pathology

Histology of kidney biopsy specimens were scored by core NEPTUNE pathologists. Global sclerosis with and without hyalinosis was reported as a percent of glomeruli counted. Arteriosclerosis and arterial hyalinosis were graded from 0 to 3+.

Covariate Variables

The NEPTUNE database was abstracted for demographic and clinical data including age, sex, race/ethnicity, glomerular disease cohort, medications, smoking exposure, follow-up time and clinical outcomes. The presence, location and severity of edema were documented by physical exam at each study visit.

Statistical Analysis

Descriptive statistics were used to compare baseline (within 45 days of biopsy) characteristics in participants with obesity versus without obesity, by adult (≥18 years) and pediatric (<18 years) subgroups. Obese and non-obese weight groups were compared using chi-square tests for categorical characteristics and Student t-tests for continuous variables. Cochran’s Q test was used to test whether the prevalence of obesity changed over time among adult and pediatric populations in participants with complete data across 5 years of follow-up.

Longitudinal linear and logistic regression using generalized estimating equations (GEE) examined the association of obesity with BP and lipids in crude and multivariable adjusted models, separately for adults and children. Adjusted models included age, sex, Black race, glomerular diagnosis, follow-up time, edema, UPC, log eGFR, and steroid use. Covariates were chosen a priori based on literature review and clinical knowledge. An unstructured correlation matrix was specified. Because the presence of significant edema might impact the calculated BMI and weight status, the interaction between obesity and edema was examined. Covariate values were log-transformed as necessary to ensure linear associations with outcomes. Results of these analyses are presented as beta coefficients (β) for continuous outcomes and odds ratios (OR) for dichotomous outcomes with corresponding 95% confidence intervals (CI).

Finally, Kaplan-Meier plots with log-rank test and Cox regression analyses with time-varying covariates were performed to test the effect of time-varying obesity on time to first complete remission, kidney failure and CKD progression, with results reported as hazard ratios (HR). Within the multivariable model for time to complete remission, time-varying variables in addition to obesity were age at visit, log eGFR and the presence of edema, along with other static covariates (considered in previous linear or logistic models using GEE) including patient sex, Black race, glomerular diagnosis and glucocorticoid treatment. For the outcome of CKD progression, a similar model was used with the exception that UPC was added as a time-varying covariate and baseline eGFR was entered as a static covariate. Given the low numbers of kidney failure events in adults, only age, sex and obesity were entered into that model. The outcomes of kidney failure and CKD progression were not analyzed in children due to the rare occurrence of the events. SPSS, Version 25 (IBM Inc.) was utilized for all analyses and a 2-sided P-value ≤0.05 was considered statistically significant.

Results

Patient Population

Data were from 541 participants enrolled in NEPTUNE, including 360 adults (baseline mean ± SD of 45.9 ± 16 years, 61% male, 23% Black) and 181 children (baseline 9.8 ± 5 years, 56% male, 41% Black) with baseline mean ± SD eGFR of 80.03 ± 39.04 ml/min/1.73m² and UPC 3.39 ± 4.4. 85% of subjects were enrolled from NEPTUNE 1 and 15% were enrolled from NEPTUNE 2. The prevalence of obesity at baseline was 43.3% (N = 156) in adults and 37.6% (N = 68) in children. The prevalence of obesity in individuals without the presence of edema at baseline was 35.2% (68/193) in adults and 32.7% (37/113) in children. Demographics and clinical characteristics of the study cohort by obesity status at baseline are described in Table 1. Systolic BP in adults and systolic BP index in children were significantly greater in participants with obesity compared to participants without obesity. Within both adult and pediatric cohorts, HDL was significantly lower in the group with obesity compared to those without obesity. However, only the adult cohort saw significant differences in edema, triglycerides, and UPC between obese and non-obese weight status at baseline.

Table 1.

Demographic and Clinical Characteristics by Baseline Weight Status.

	Adult			Pediatric
N (%) or mean ± SD	Obese N = 156	Non-Obese N = 204	p-value	Obese N = 68	Non-Obese N = 113	p-value
Age (years)	46.0 ± 14.2	45.8 ± 17.4	0.9	9.6 ± 4.9	10.0 ± 4.8	0.6
Male	91 (58.3)	130 (63.7)	0.3	37 (54.4)	65 (57.5)	0.7
Black	41 (26.3)	41 (20.1)	0.2	28 (41.2)	46 (40.7)	0.9
Hispanic	39 (25.0)	35 (17.2)	0.09	13 (19.1)	27 (23.9)	0.3
BMI (kg/m2)	36.7 ± 7.5	25.1 ± 3.1	<0.001	-	-	-
BMI percentile	-	-	-	98.2 ± 1.4	65.3 ± 28.4	<0.001
Cohort: MCD	23 (14.7)	41 (20.1)	0.2	35 (51.5)	59 (52.2)	0.4
MN	44 (28.2)	46 (22.5)		1 (1.5)	0 (0)
FSGS	55 (35.3)	63 (30.9)		23 (33.8)	32 (28.3)
IgA	16 (10.3)	32 (15.7)		4 (5.9)	5 (4.4)
Follow-up (months)	31.9 ± 20.3	33.7 ± 19.5	0.4	37.8 ± 19.6	35.6 ± 19.5	0.5
Edema	88 (56.4)	79 (38.7)	0.001	31 (45.6)	36 (31.9)	0.07
Edema Location:			0.2			0.9
Face/Periorbital	1 (0.6)	6 (2.9)		9 (13.2)	8 (7.1)
Lower Extremity	65 (41.7)	51 (25)		12 (17.6)	16 (14.2)
Sacral	10 (6.4)	9 (4.4)		1 (1.5)	2 (1.8)
Anasarca	12 (7.7)	13 (16.4)		9 (13.23)	10 (8.8)
RAAS use	95 (60.9)	104 (51.0)	0.06	21 (30.9)	34 (30.1)	0.9
Steroid use	32 (20.5)	44 (21.6)	0.8	36 (52.9)	65 (57.5)	0.6
CNI use	2 (1.3)	8 (3.9)	0.1	17 (25.0)	19 (16.8)	0.2
Smoker	17 (10.9)	17 (8.3)	0.4	0 (0)	1 (0.9)	0.4
SBP (mmHg) for adults/SBP index for children*	127.6 ± 16.7	123.4 ± 18.2	0.02	0.96 ± 0.11	0.92 ± 0.1	0.03
DBP (mmHg) for adults/DBP index for children*	77.7 ± 10.7	75.9 ± 12.1	0.1	0.93±0.17	0.92 ± 0.16	0.3
Total Cholesterol (mg/dl)	257.5 ± 106.4	268.8 ± 100.9	0.3	279.8 ± 132.01	305.4 ± 138.7	0.2
HDL (mg/dl)	60.6 ± 27.5	72.5 ± 30.0	<0.001	70.9 ± 30.3	81.6 ± 26.3	0.02
LDL (mg/dl)	150.4 ± 87.2	160.3 ± 84.5	0.3	165.8 ± 104.2	189.5 ± 110.9	0.2
Triglycerides (mg/dl)	232.4 ± 161.6	179.9 ± 127.4	0.001	215.9 ± 184.7	171.2 ± 127	0.07
UPC	4.43 ± 5.39	3.03 ± 3.57	0.003	3.51 ± 5.27	2.49 ± 3.26	0.1
eGFR	66.8 ± 30.3	70.2 ± 31.9	0.3	99.9 ± 54.1	104.6 ± 35.0	0.5

SD – standard deviation; BMI – body mass index; MCD – minimal change disease; MN – membranous nephropathy; FSGS – focal segmental glomerulosclerosis; IgA – immunoglobulin A; RAAS – renin– angiotensin–aldosterone system; CNI – calcineurin inhibitor; SBP – systolic BP; DBP – diastolic BP; HDL – high-density lipoproteins; LDL – low-density lipoproteins; UPC – urine protein creatinine ratio; eGFR – estimated glomerular filtration rate

Obesity and Edema

The prevalence of obesity overall in adults and children did not significantly change over study follow-up (Cochran’s Q = 7.43, p = 0.4); see Figure 1. Figure 2 demonstrates the prevalence of non-obesity, obesity with edema and obesity without edema in adults and children over the study period. In adults, while 24.4% of participants had obesity and edema at baseline, this number declined steadily to just 5.5% at the 60-month visit. In conjunction, the proportion of individuals with obesity and no edema significantly increased (Cochran’s Q = 12.96, p = 0.04). In the pediatric cohort, the presence of edema also decreased over time. At baseline, 17.1% of participants had obesity and edema and this proportion declined to 3.4% at the 60-month visit. However, the proportion of individuals with obesity and no edema did not significantly change over time (Cochran’s Q = 9.75, p = 0.1).

Figure 1.

Prevalence of obesity among children and adults with proteinuric glomerulopathies enrolled in NEPTUNE. The proportion with obesity did not vary significantly over time (Cochran’s Q = 7.43, p = 0.4).

Figure 2.

Prevalence of obesity and edema in adults (left) and children (right) with proteinuric glomerulopathies over time. Cochran’s Q indicated significant increase in obese nonedematous adults over time (Q = 12.96, p = 0.04). There was no significant change in obese nonedematous children over time (Q = 9.75, p = 0.1).

Obesity and Cardiovascular Measures

At baseline, the adult group with obesity had a significantly higher proportion with dyslipidemia compared to the group without obesity. There was no significant difference in baseline HTN between groups in adults or children (Table 2). In linear and logistic GEE models, the interaction between obesity and edema was not significant and was thus not retained in the final models. In adults, obesity was independently associated with higher systolic BP (β=6.49, 95% CI: 2.41,10.56, p=0.002), triglycerides (β=41.92, 95% CI: 17.12,66.71, p=0.001), and odds of dyslipidemia (OR=1.74, 95% CI: 1.30,2.32, p<0.001). Obesity was also independently associated with lower HDL (β=−6.92, 95% CI: −9.32,−4.51, p<0.001; Table 3). Within the pediatric population, adjusted GEE models showed a significant association between obesity and greater systolic BP index (β=0.04, 95% CI: 0.02,0.06, p<0.001) and hypertension (OR=1.43, 95% CI: 1.04,1.98, p=0.03).

Table 2.

Cardiovascular Measures, Kidney Disease Outcomes and Pathology by Baseline Weight Status.

	Adult			Pediatric
N (%) or mean ± SD	Obese N = 156	Non-Obese N = 204	p-value	Obese N = 68	Non-Obese N = 113	p-value
Cardiovascular
Hypertensive BP Status at Baseline	52 (33.3)	60 (29.4)	0.4	32 (47.1)	44 (38.9)	0.3
Dyslipidemia at Baseline	124 (79.5)	136 (66.7)	0.007	54 (79.4)	73 (64.6)	0.2
Kidney Disease*
CKD Progression	39 (25.0)	43 (21.1)	–	16 (23.5)	20 (17.7)	–
Kidney failure	14 (9.0)	16 (7.8)	–	0 (0)	2 (1.8)	–
Complete Remission Ever	72 (46.2)	112 (54.9)	–	46 (67.6)	81 (71.7)	–
Pathology
Global Sclerosis without hyalinosis	N=47 11.7 ± 13.6	N=76 10.8 ± 14.3	0.7	N=41 2.7 ± 8.3	N=64 1.5 ± 4.3	0.3
Global Sclerosis with hyalinosis	N=47 4.5 ± 9.3	N=76 2.8 ± 7.1	0.3	N=41 0.76 ± 4.2	N=64 0.24 ± 1.3	0.4
Arteriosclerosis (0-3)	N=84 0.84 ± 0.92	N=122 0.91 ± 0.94	0.6	N=48 0.06 ± 0.22	N=74 0.06 ± 0.22	0.9
Arterial Hyalinosis (0-3)	N=93 0.35 ± 0.54	N=128 0.3 ± 0.51	0.5	N=49 0.03 ± 0.16	N=76 0.01 ± 0.06	0.2

^*The numbers for kidney disease outcomes are not adjusted for follow-up time, and should not be statistically compared.

Table 3.

Association of Obesity with Outcome Measures in Adult and Pediatric Participants with Glomerular Disease. Results Shown for Crude and Adjusted Linear Longitudinal Models, Logistic Longitudinal Models, and Cox Regression with Time-Varying Covariates.

	Adult		Pediatric
Outcome	Crude	Adjusted*	Crude	Adjusted*
Linear Longitudinal Models (using GEE) with Continuous Outcomes(β, 95th% CI, p-value)
SBP (mm Hg)	3.20 (1.10, 5.29) p=0.003	6.49 (2.41, 10.56) p=0.002	------	------
SBP index	------	------	0.02 (0.01, 0.04) p=0.001	0.04 (0.02, 0.06) p<0.001
DBP	1.40 (−0.04, 2.85) p=0.06	2.31 (−0.81, 5.43) p=0.2		------
DBP index	------	------	0.05 (0.01, 0.1) p=0.01	0.02 (−0.03, 0.08) p=0.4
Triglycerides	12.84 (−21.70, 47.37) p=0.5	41.92 (17.12, 66.71) p=0.001	42.03 (−21.27, 105.33) p=0.2	24.77 (−6.22, 55.76) p=0.1
HDL	−6.95 (−10.36, −3.54) p<0.001	−6.92 (−9.32, −4.51) p<0.001	−0.47 (−12.27, 11.33) p=0.9	8.91 (−0.26, 18.09) p=0.06
LDL	1.22 (−8.89, 11.34) p=0.8	−6.26 (−15.02, 2.49) p=0.2	13.13 (−8.36, 34.61) p=0.2	12.71 (−18.44, 43.86) p=0.4
Total Cholesterol	0.12 (−11.09, 11.35) p=0.9	−8.51 (−17.46, 0.45) p=0.06	10.66 (−9.58, 30.89) p=0.3	14.63 (−28.96, 58.21) p=0.5
Logistic Longitudinal Models (using GEE) with Dichotomous Outcome (OR, 95% CI, p-value)
Hypertensive BP Status	1.26 (1.02, 1.54) p=0.03	1.21 (0.93, 1.58) p=0.2	2.13 (1.57, 2.89) p<0.001	1.43 (1.04, 1.98) p=0.03
Dyslipidemia	1.42 (0.98, 2.05) p=0.07	1.74 (1.30, 2.32) p<0.001	3.67 (1.32, 10.23) p=0.01	2.78 (0.65, 11.79) p=0.2
Cox Regression with Time-Varying Covariates (HR, 95th% CI, p-value)**
Time to CKD Progression	1.31 (0.97, 1.77) p=0.08	0.80 (0.67, 1.06) p=0.9	------	------
Time to Kidney Failure	0.66 (0.38, 1.13) p=0.1	0.90 (0.66, 1.20) p=0.5	------	------
Time to Complete	0.60 (0.51, 0.71) p<0.001	0.80 (0.69, 0.88) p<0.001	0.59 (0.49, 0.73) p<0.001	0.72 (0.61, 0.84) p<0.001
Remission of Proteinuria

GEE – generalized estimating equations; OR – odds ratio; HR – hazard ratio; CI – confidence interval; SBP – systolic blood pressure; DBP – diastolic blood pressure; HDL – high density lipoprotein; LDL – low density lipoprotein.

Definitions: Hypertensive BP Status = average BP in the hypertensive range at each study visit or if a clinical diagnosis of hypertension was documented in their medical record. Dyslipidemia = HDL <40 mg/dL; non-HDL ≥145 mg/dl (<18 years), ≥160 mg/dl (≥18 years); or triglycerides ≥100 mg/dL (0-9 years), ≥130 mg/dL (10-17 years), or ≥150 mg/dL (≥18 years) at each study visit. CKD Progression = having both an estimated glomerular filtration rate (eGFR) decline by 40% and an eGFR <90 ml/min/1.73m², or developing End Stage Renal Disease (ESRD). Kidney Failure = either two consecutive study visits with eGFR values <15 ml/min/1.73m² or ESRD. Time to Complete Remission = urine protein to creatinine ratio (UPC) <0.3 g/g at any time point after diagnosis

^*Linear and logistic longitudinal models using GEE included obesity, age, sex, Black race, glomerular diagnosis, follow-up time, edema, UPC, log eGFR, and steroid use. Cox regression models adjusted for time-varying variables of obesity, age at visit, log eGFR and edema, and static covariates of sex, Black race, glomerular diagnosis and steroids for the outcome of Time to Complete Remission. For the outcome of CKD progression, time varying UPC was included and static baseline eGFR was used. Given the low numbers of events in adults, only age, sex and obesity were entered into the model for kidney failure.

^**CKD Progression and Kidney Failure not analyzed in children due to the low number of events.

Obesity and Kidney Disease Outcomes

During the course of the study, 274 (73.9%) adults and 144 (75%) children achieved partial or complete remission. In adults, Cox regression models with time-varying obesity found no significant relationships between obesity and the hazard of kidney failure or CKD progression. However, obesity was associated with a lower hazard of ever achieving complete remission (HR=0.80, 95% CI: 0.69,0.88, p<0.001; Table 3, Figure 3). Similarly, children with obesity had a lower hazard of ever reaching complete remission (HR=0.72, 95% CI: 0.61,0.84, p<0.001; Table 3, Figure 4) compared to those without obesity. Log-rank test also found significant differences in the proportion that reached remission in both adults and children (p<0.001). Very few children reached the outcomes of kidney failure (N = 2) or CKD progression (N = 36).

Figure 3.

Kaplan-Meier plot of complete remission by baseline obesity status in adults. There was a significant difference in the proportion that reached remission (log-rank test p <0.001).

Figure 4.

Kaplan-Meier plot of complete remission by baseline obesity status in children. There was a significant difference in the proportion that reached remission (log-rank test p <0.001).

Obesity and Pathology Markers

There were no significant differences in global sclerosis with and without hyalinosis, arteriosclerosis or arterial hyalinosis in adults or children with and without obesity at baseline (Table 2).

Discussion

In this large, multi-center cohort of adults and children with primary glomerulopathies, 156 (43.3%) adults and 68 (37.6%) children were found to have obesity at the time of enrollment. Obesity at baseline was associated with more frequent edema, higher triglycerides, and a higher UPC in adults, and with a higher systolic BP and lower HDL in both adults and children. There were no significant differences in kidney histology (glomerulosclerosis, arterial hyalinosis and arteriosclerosis) by weight status. By multivariable analyses, obesity in adults was independently associated with higher systolic BP, higher triglycerides, lower HDL and greater odds of dyslipidemia. Within the pediatric population, obesity was independently associated with higher systolic BP index and hypertension. Finally, in both adults and children, obesity was associated with significantly lower odds of ever reaching complete remission of glomerular disease.

Obesity is an established risk factor for CVD and contributes to the development and progression of CKD. In this observational cohort, obesity was highly prevalent; even when considering those without edema, the baseline prevalence of obesity still included 68 (35.2%) adults and 37 (32.7%) children. Recent US public health data points to comparable obesity levels among the general population of adults, at 39.8%, but much lower among the general population of children, at 18.5% [1]. Although the overall prevalence of obesity as determined by BMI did not change over the 60-month period, there may have been an actual increase in true obesity over time: the steady decline in edema over time did not result in a decrease in obesity prevalence. A plausible explanation for this may be true weight gain as a side effect of treatment medications such as glucocorticoids.

There are several mechanisms by which obesity may affect kidney function. Increased fat mass contributes to mesangial expansion [9] and elevated renal metabolic load [10], promoting glomerular hyperfiltration and hypertrophy [12, 11]. These changes may trigger downstream biochemical cascades, such as the renin-angiotensin and TGF-β systems, exacerbating kidney damage [14]. Moreover, autopsies of both adult and child kidneys have revealed increased kidney weight among those with obesity, possibly related to the compensatory hypertrophy of individual glomeruli [26]. No differences in kidney histology were found in the current study, although glomerular size was not available for analysis at the time of this publication. An independent analysis of kidney biopsies of obese versus non-obese diabetic nephropathy patients revealed heavier lipid deposition and increased intercellular lipid droplets in the renal specimen from those with obesity [27]. Down regulation of fatty acid β-oxidation pathways, including PPAR-α, carnitine palmitoyltransferase 1, and acyl-CoA oxidase, were also observed in these kidneys. There are also changes starting at the transcriptomic level within the glomerulus with obesity. A small study of adults with obesity-related glomerulopathy found that there was differential expression of genes, extracted as RNA from renal glomeruli, involved in the regulation of lipid metabolism, inflammation and insulin that may contribute to the pathogenesis of glomerular damage [28].

Unsurprisingly, obesity was associated with higher systolic BP/BP index in both adults and children, at baseline and in adjusted models over time. Obesity has been implicated in the development of CVD, specifically heart failure and coronary atherosclerosis. Physiologically, the presence of excess adipose tissue leads to hyperproduction of pro-inflammatory cytokines, which may impair cardiac ventricular function and promote atherosclerotic plaques [29]. One study found that a 10 kg increase in body weight resulted in a 12% increased chance of developing coronary artery disease, along with a 3 mmHg increase in SBP [30]. Although the general relationship between obesity and CVD risk is established, it is important to ascertain the direction and magnitude of these associations in specific high-risk patient groups such as those with glomerular diseases.

Importantly, we found that obesity was associated with a significantly lower chance of ever reaching complete remission from glomerular diseases in both adults and children. Obesity was not associated with a greater hazard of reaching CKD progression or kidney failure in adults, although the frequency of these outcomes might have been too low, or duration of follow-up too short, to identify such associations in children. The former result is however consistent with a European single-center retrospective cohort study of adults with primary glomerular diseases (excluding MCD), which found that individuals with obesity were not more likely to progress to a composite endpoint of CKD Stage 5 or renal replacement therapy [16]. While this study went on to recommend against focusing on weight loss for primary glomerular disease, it did not consider the significant effect of obesity on complete remission.

We propose a couple of theories as to why individuals with obesity saw lower odds of complete remission. For one, as mentioned previously, the direct effect of obesity on the glomerulus may alter glomerular structure thereby modifying the disease process and the ability to achieve remission. Secondly, patients with obesity may metabolize certain immunosuppressant medications differently, impacting their therapeutic benefit [31, 32]. A review of corticosteroid pharmacokinetic abnormalities in patients with asthma and obesity found a significant inverse relationship between plasma prednisolone and TNFα mRNA levels, but this relationship lost significance when adjusting for BMI, implying that BMI may affect the drug’s metabolic clearance rate [32]. In fact, adults with obesity have enhanced 5α-reductase type 1 activity in the liver, as evidenced by increased excretions of cortisol as 5α-reduced metabolites when compared with adults who are not obese. This change in corticosteroid metabolism has also been replicated in studies of obese versus lean Zucker rats [31]. Similarly, calcineurin inhibitors (CNI) may also be metabolized differently in the presence of obesity [33]. However, CNI dosing can be guided by steady-state trough levels, whereas this type of adjustment is not performed with steroids. As such, adults are given standard dosages of steroids and children are given weight-dependent dosages without taking fat mass into account. Therefore, it is plausible that a higher metabolic clearance rate of corticosteroids in individuals with obesity may lead to decreased drug response, thereby resulting in lower odds of reaching complete remission.

Our study has several limitations. Firstly, the definition of obesity used in this study based on BMI could have over-estimated the true prevalence of obesity due to edema. As such, an individual with fluid overload could have an overstated BMI, resulting in a misclassification of obesity status [34]. However, we adjusted for the presence of edema in models and considered the interaction between obesity and edema, which was not statistically significant. The presence of edema also decreased over time and was adjusted for in longitudinal models. Additionally, since this was an observational cohort, the confounding of anti-hypertensive medications, lipid lowering medications and immunosuppression medications use varied throughout the study period and may have had an effect on the cardiovascular and kidney disease outcomes of interest. Given the observational nature of the study, the relationship between obesity and outcomes may not necessarily be causal, and future studies examining the effects of weight loss on cardiovascular risk profile and kidney disease outcomes in this specific patient population are required. Furthermore, although the average follow-up time in this cohort was close to three years, it is possible that participants were not followed for a long enough time period to see the effect of obesity on certain outcomes, especially those related to CKD progression and renal failure. Lastly, all NEPTUNE study sites are academic centers, and generalizability to other practice settings remains unknown. This study does have several notable strengths. It is a large cohort with repeated measures of individuals with biopsy-proven diagnoses of specific proteinuric glomerulopathies.

In this prospective cohort study of patients with incident, biopsy-proven, idiopathic glomerular diseases, obesity was associated with a higher CVD risk profile and a lower chance of complete remission of proteinuria. Obesity is a modifiable risk factor that therefore presents itself as a high-priority therapeutic target to improve short and long-term clinical outcomes in this high-risk patient group. Future studies aimed at elucidating the mechanisms by which obesity affects glomerular disease remission could guide future drug discovery.