Comparative Efficacy of Nonsteroid Immunosuppressive Medications in Childhood Nephrotic Syndrome

Cal H Robinson; Nowrin Aman; Tonny Banh; Josefina Brooke; Rahul Chanchlani; Brian H Cuthbertson; Vaneet Dhillon; Eddy Fan; Valerie Langlois; Leo Levin; Christoph Licht; Ashlene McKay; Damien Noone; Rachel Pearl; Seetha Radhakrishnan; Veronique Rowley; Chia Wei Teoh; Jovanka Vasilevska-Ristovska; Anna Heath; Rulan S Parekh

doi:10.1001/jamapediatrics.2024.5286

. 2025 Jan 27;179(3):321–331. doi: 10.1001/jamapediatrics.2024.5286

Comparative Efficacy of Nonsteroid Immunosuppressive Medications in Childhood Nephrotic Syndrome

Cal H Robinson ^1,^2,^3,^✉, Nowrin Aman ³, Tonny Banh ³, Josefina Brooke ¹, Rahul Chanchlani ^4,⁵, Brian H Cuthbertson ^6,^7,⁸, Vaneet Dhillon ¹, Eddy Fan ⁶, Valerie Langlois ⁹, Leo Levin ^1,², Christoph Licht ^1,^2,¹⁰, Ashlene McKay ^1,², Damien Noone ^1,², Rachel Pearl ^1,^2,¹¹, Seetha Radhakrishnan ^1,², Veronique Rowley ³, Chia Wei Teoh ^1,², Jovanka Vasilevska-Ristovska ³, Anna Heath ^3,^12,¹³, Rulan S Parekh ^1,^2,^3,¹⁴

¹Division of Nephrology, The Hospital for Sick Children, Toronto, Ontario, Canada

²Department of Paediatrics, The University of Toronto, Toronto, Ontario, Canada

³Child Health Evaluative Sciences, Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada

⁴Division of Nephrology, Department of Pediatrics, McMaster Children’s Hospital, Hamilton, Ontario, Canada

⁵Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, Ontario, Canada

⁶Interdepartmental Division of Critical Care Medicine, The University of Toronto, Ontario, Canada

⁷Sunnybrook Health Sciences Centre Research Institute, Toronto, Ontario, Canada

⁸Department of Critical Care Medicine, Sunnybrook Health Sciences Centres, Toronto, Ontario, Canada

⁹Division of Nephrology, Department of Paediatrics, Montreal Children’s Hospital, Montreal, Quebec, Canada

¹⁰Program in Cell Biology, Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada

¹¹Division of Nephrology, William Osler Health Systems, Brampton, Ontario, Canada

¹²Division of Biostatistics, Dalla Lana School of Public Health, The University of Toronto, Toronto, Ontario, Canada

¹³Department of Statistical Science, University College London, London, United Kingdom

¹⁴Department of Medicine, Women’s College Hospital, Toronto, Ontario, Canada

Accepted for Publication: September 16, 2024.

Published Online: January 27, 2025. doi:10.1001/jamapediatrics.2024.5286

^✉

Corresponding Author: Cal H. Robinson, MBChB, SickKids Research Institute, Peter Gilgan Centre for Research and Learning, 686 Bay St, Toronto, ON M5G 0A4, Canada (cal.robinson@sickkids.ca).

Author Contributions: Drs Robinson and Parekh had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Heath and Parekh were co-senior authors.

Concept and design: Robinson, Cuthbertson, Dhillon, Fan, Licht, Noone, Teoh, Parekh.

Acquisition, analysis, or interpretation of data: Robinson, Aman, Banh, Brooke, Chanchlani, Cuthbertson, Fan, Langlois, Levin, McKay, Pearl, Radhakrishnan, Rowley, Teoh, Vasilevska-Ristovska, Heath, Parekh.

Drafting of the manuscript: Robinson, Cuthbertson.

Critical review of the manuscript for important intellectual content: Aman, Banh, Brooke, Chanchlani, Cuthbertson, Dhillon, Fan, Langlois, Levin, Licht, McKay, Noone, Pearl, Radhakrishnan, Rowley, Teoh, Vasilevska-Ristovska, Heath, Parekh.

Statistical analysis: Robinson, Cuthbertson, Radhakrishnan, Rowley, Heath, Parekh.

Obtained funding: Parekh.

Administrative, technical, or material support: Aman, Banh, Brooke, Chanchlani, Cuthbertson, Dhillon, Pearl, Teoh, Vasilevska-Ristovska, Parekh.

Supervision: Banh, Chanchlani, Cuthbertson, Fan, McKay, Noone, Radhakrishnan, Teoh, Heath, Parekh.

Conflict of Interest Disclosures: Dr Fan reported personal fees from Baxter, Getinge, Inspira, and Vasomune outside the submitted work. Dr Parekh reported a consulting agreement with Vertex and Sanofi for clinical trials in nephrotic syndrome and chronic kidney disease. Dr Robinson received salary support from the Cure Glomerulopathy consortium Career Development Fellowship Program, the SickKids Clinician-Scientist Training Program, and the Canadian Institutes of Health Research Fellowship Program. Dr Parekh was supported by the Women’s College Hospital Research Institute F.M. Hill Chair in Health System Solutions. No other disclosures were reported.

Data Sharing Statement: See Supplement 2.

Additional Contributions: We thank the participants and their families for their time and effort, as well as the nurses and staff from the Nephrology Clinic at The Hospital for Sick Children, William Osler Health System (Brampton Civic Hospital and Peel Memorial Centre), Scarborough Health Network (Centenary Hospital), and McMaster Children’s Hospital.

^✉

Corresponding author.

PMCID: PMC11773402 PMID: 39869322

This study evaluates the comparative effectiveness of cyclophosphamide vs calcineurin inhibitors (tacrolimus or cyclosporine) for childhood nephrotic syndrome relapse prevention.

Key Points

Question

Are calcineurin inhibitors more effective than cyclophosphamide at preventing childhood steroid-sensitive nephrotic syndrome relapses?

Findings

This observational study emulated a pragmatic trial comparing outcomes in cyclophosphamide vs calcineurin inhibitors among a cohort of 578 children with nephrotic syndrome. During a median 5.5-year follow-up, there was no significant difference in time to relapse, relapse rates, subsequent immunosuppression use, or kidney function between medications and calcineurin inhibitor treatment was associated with more hospitalizations and intravenous albumin use.

Meaning

In this study, cyclophosphamide was a preferred first-line nonsteroid immunosuppressive medication, since it is lower cost, shorter duration, and more accessible globally than calcineurin inhibitors.

Abstract

Importance

Cyclophosphamide and calcineurin inhibitors are the most used nonsteroid immunosuppressive medications globally for children with various chronic inflammatory conditions. Their comparative effectiveness remains uncertain, leading to worldwide practice variation. Nephrotic syndrome is the most common kidney disease managed by pediatricians globally and suboptimal treatment is associated with high morbidity.

Objective

To evaluate the comparative effectiveness of cyclophosphamide vs calcineurin inhibitors (tacrolimus or cyclosporine) for childhood nephrotic syndrome relapse prevention.

Design, Setting, and Participants

Using target trial emulation methods, the study team emulated a pragmatic, open-label clinical trial using available data from the Insight Into Nephrotic Syndrome: Investigating Genes, Health, and Therapeutics (INSIGHT) study. INSIGHT is a multicenter, prospective cohort study in the Greater Toronto Area, Canada. Participants included children (1 to 18 years) with steroid-sensitive nephrotic syndrome diagnosed between 1996 and 2019 from the Greater Toronto Area, who initiated cyclophosphamide or a calcineurin inhibitor treatment. Data analysis was performed in 2024.

Exposures

Incident cyclophosphamide or calcineurin inhibitor treatment. Randomization was emulated by overlap weighting of propensity scores for treatment assignment.

Main Outcomes

The primary outcome was time to relapse, analyzed by weighted Kaplan-Meier and Cox proportional hazards models. Secondary outcomes included relapse rates, subsequent immunosuppression, kidney function, hypertension, adverse events, and quality of life.

Results

Of 578 children (median age at diagnosis, 3.7 [IQR, 2.8-6.0] years; 371 male [64%] and 207 female [36%]), 252 initiated cyclophosphamide, 131 initiated calcineurin inhibitors, and 87 sequentially initiated both medications. Baseline characteristics were well balanced after propensity score weighting. During median 5.5-year (quarter 1 to quarter 3, 2.5-9.2) follow-up, there was no significant difference in time to relapse between calcineurin inhibitor vs cyclophosphamide treatment (hazard ratio [HR], 1.25; 95% CI, 0.84-1.87). Relapses were more common after calcineurin inhibitor treatment than cyclophosphamide (85% vs 73%) in the weighted cohorts, but not statistically significant. There were also no significant differences in subsequent relapse rates, nonsteroid immunosuppression use, or kidney function between medications. Calcineurin inhibitor treatment was associated with more hospitalizations (HR, 1.83; 95% CI, 1.14-2.92) and intravenous albumin use (HR, 2.81; 95% CI, 1.65-4.81).

Conclusions and Relevance

In this study, there was no evidence of difference in time to relapse after cyclophosphamide and calcineurin inhibitor treatment in children with nephrotic syndrome. Cyclophosphamide treatment is shorter in duration and more accessible globally than calcineurin inhibitors.

Introduction

Nephrotic syndrome is the most common kidney disease managed by pediatricians worldwide and is associated with high disease-related and treatment-related morbidity.^1,2,3,4,5 Nephrotic syndrome is treated initially and at each relapse with steroid immunosuppression. Currently, approximately 90% of children are steroid sensitive, but most relapse and receive repeated steroid courses.^1,2 Frequent relapses cause steroid-related complications and poor quality of life.^1,5,6,7 Half of all children with nephrotic syndrome receive nonsteroid immunosuppressive medications to prevent relapses.⁸ Cyclophosphamide and tacrolimus (a calcineurin inhibitor) are recommended by the most recent Kidney Disease: Improving Global Outcomes and International Pediatric Nephrology Association guidelines.^9,10 However, no randomized clinical trial (RCT) has ever directly compared them.¹¹ Instead, data are extrapolated from case reports, other diseases, network meta-analysis, and historical medications (chlorambucil and cyclosporine).^11,12 These data suggest that relapse rates are similar within 1 to 2 years of starting either treatment but more than 50% of children relapse after stopping calcineurin inhibitors.^13,14,15,16

The comparative effectiveness of cyclophosphamide and calcineurin inhibitors in childhood nephrotic syndrome remains unclear, which is a key knowledge gap.⁹ Physician prescribing patterns vary worldwide.^17,18,19 This is a global health concern, since higher drug cost, limited access, and laboratory surveillance burden are major barriers to calcineurin inhibitor use in resource-limited health care systems.²⁰ Cyclophosphamide is typically administered daily for 8 to 12 weeks, whereas calcineurin inhibitors are given for longer than 1 to 2 years with regular laboratory and therapeutic drug monitoring. Despite concerns about long-term infertility and cancer risks after cyclophosphamide treatment,^21,22,23,24 it remains one of the most commonly prescribed immunosuppressive medications worldwide.^{5,17,18,25,26,27}

Although an RCT would be the ideal study design to compare cyclophosphamide against calcineurin inhibitors, it is infeasible based on the anticipated sample size, costs, and patient and physician medication preferences. Target trial emulation is an alternative approach, which applies RCT design principles to allow for causal inferences from observational data by addressing common observational research biases and measured confounding.^28,29,30 To compare the effectiveness of cyclophosphamide vs calcineurin inhibitors for nephrotic syndrome relapse prevention, we emulated a hypothetical pragmatic, open-label RCT using observational data available from a large, prospective childhood nephrotic syndrome cohort.

Methods

Study Design

We performed a retrospective analysis of data from the Insight Into Nephrotic Syndrome: Investigating Genes, Health, and Therapeutics (INSIGHT) study (eTable 1 in Supplement 1).³¹ Full INSIGHT methods were previously published.³¹ Briefly, INSIGHT is a prospective longitudinal childhood nephrotic syndrome cohort enrolling children diagnosed between 1996 and 2019 from the Greater Toronto and Hamilton Area, Canada. Ethics approval was obtained at each clinical center with written informed consent and assent from all study participants. Nephrotic syndrome is defined using Kidney Disease: Improving Global Outcomes criteria (ie, the presence of proteinuria [40 or more mg/m² per hour, urine protein/creatinine ratio 200 mg/mL or more, or 3 positive urine dipstick protein test results], hypoalbuminemia [less than 25 g/L], and edema).⁹ This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines (eAppendix in Supplement 1).³²

Population

We included all INSIGHT participants (1 to 18 years old) diagnosed with nephrotic syndrome between 1996 and 2019 who initiated cyclophosphamide or calcineurin inhibitor (tacrolimus or cyclosporine) treatment during follow-up. No formal power calculation was performed since this study used established observational data. Children were excluded if they had steroid-resistant disease, genetic or secondary nephrotic syndrome (eg, lupus nephritis), or biopsy-proven membranous nephropathy or membranoproliferative glomerulonephritis.³¹

Outcomes

Our primary outcome was time to relapse after cyclophosphamide or calcineurin inhibitor initiation. Relapses were defined as recurrent nephrotic range proteinuria (ie, 3 or more positive urine dipstick protein test results) for 3 or more consecutive days.⁹ Relapses were detected by home urine monitoring and typically confirmed by laboratory urine testing. Outcome follow-up began at medication initiation. Participants were followed up until transition to adult services, nephrology clinic discharge (after 4 to 5 relapse-free years), loss to follow-up, or administrative censoring on December 31, 2019. Secondary outcomes were total relapse count and relapse occurrence (yes/no) by year 1, year 2, and year 5 after medication initiation, change in estimated glomerular filtration rate (eGFR) (Chronic Kidney Disease in Children U25 formula)³³ from baseline to last follow-up measurement, incident chronic kidney disease (stage 3 positive; eGFR measurement less than 60 mL/min/1.73m² among children with a baseline eGFR more than 60 mL/min/1.73 m²),³⁴ subsequent immunosuppressive medications, hypertensive blood pressure (stage 1 to 2 hypertension by American Academy of Pediatrics 2017 classification),³⁵ antihypertensive medication initiation, adverse events (hospitalizations, intravenous albumin use, and infections), and change in quality-of-life score (PedsQL)³⁶ from baseline to first follow-up measurement.

Exposure and Comparator

We compared the effectiveness of incident cyclophosphamide vs calcineurin inhibitor treatment, defined by medication initiation (ie, modified intention to treat). Participants who received both cyclophosphamide and tacrolimus sequentially at different time points were included in both treatment arms, with separate propensity scores calculated based on updated baseline covariates. At INSIGHT centers, oral cyclophosphamide was typically administered at a dose of 2 mg/kg daily for 8 to 12 weeks (maximum dose = 168 mg/kg), adjusted based on weekly blood cell counts. The typical initial tacrolimus trough level target was 3 to 5 ng/mL, which was increased to 5 to 7 ng/mL if relapses continued. Target cyclosporine levels were typically 100 to 200 ng/mL. If effective, calcineurin inhibitors were continued for 2 years before attempted weaning.

Baseline Characteristics

Demographic covariates included age, sex, self-reported ethnicity, immigration status, geographic area and rural residence (by postal code), and parental education (highest level achieved by either parent). Clinical covariates were center (academic vs community hospital), baseline body mass index (calculated as weight in kilograms divided by height in meters squared), hospitalization at diagnosis, nephrotic syndrome classification premedication initiation (ie, steroid dependent, frequently relapsing, or infrequently relapsing), time to medication initiation, and time to first relapse (from the diagnosis date), number of prior relapses (premedication initiation), baseline blood pressure classification,³⁵ baseline eGFR,³³ atopy history, and family history of hypertension, diabetes, or kidney disease. Medication covariates were treatment era, initial prednisone treatment duration, maintenance prednisone use (within 1 year of premedication initiation), prior nonsteroid immunosuppressive treatment, and prior antihypertensive medication use. These covariates were selected based on reported or theoretical associations with immunosuppressive treatment selection or nephrotic syndrome relapse. Baseline covariates with missing data were handled using single imputation of mean values (continuous variables) or missing categories (categorical variables) to avoid exclusion due to missing data when constructing propensity scores.

Statistical Analysis

Using target trial methods (eTable 1 in Supplement 1), randomization was emulated by overlap weighting propensity scores for treatment assignment (cyclophosphamide vs calcineurin inhibitor). Propensity scores were estimated by logistic regression, including all baseline demographic, clinical, and medication covariates. Overlap weighting then reweighted participants by the probability of counterfactual treatment (eg, 1 minus the propensity score for treated individuals).³⁷ Compared with inverse probability weighting, overlap weighting is more robust to extreme weights, more precise, and estimates average treatment effect for individuals at clinical equipoise.^37,38,39 We assessed population overlap using propensity score distribution plots. Covariate balance preweighting and postweighting were evaluated using descriptive statistics, standardized differences (≥0.1 is substantial^40,41), and Love plots.⁴²

Weighted Kaplan-Meier estimators and unadjusted Cox proportional hazards models (with robust, sandwich-type variance estimation) were used to estimate the association between treatment assignment and time to relapse. For time-to-event secondary outcomes, we used weighted Kaplan-Meier estimators and unadjusted Cox proportional hazards models. Relapse count, relapse occurrence, and change in eGFR and PedsQL scores were analyzed using weighted, unadjusted negative binomial, logistic, and linear regression, respectively. We evaluated for residual confounding using E-values and control outcomes.^28,43 Negative control outcomes (inhaler and vitamin/supplement use) were not expected to be correlated with treatment, whereas our positive control outcome (serum creatinine measurement) was expected to be more common after calcineurin inhibitor treatment. All statistical analyses were performed using R version 4.3.1 (The R Project). Propensity score methods were implemented using the WeightIt package version 1.1.0 (The R Project).⁴⁴

We evaluated for treatment-effect heterogeneity using subgroup analyses and interaction terms for: (1) age (younger than 8 years or 8 years or older), (2) sex, (3) ethnicity, (4) nephrotic syndrome classification, and (5) clinical center type. We calculated subgroup balancing propensity scores to achieve population overlap and covariate balance for each stratified analysis.⁴⁵ Sensitivity analyses were used to test the robustness of our results, by (1) restricting to the first use of either cyclophosphamide vs calcineurin inhibitor, (2) restricting to cyclophosphamide vs tacrolimus initiators after January 1, 2004 (ie, era of tacrolimus availability), (3) restricting to adherent individuals (ie, a per protocol analogue; treatment adherence was defined as 2 years or longer of consecutive calcineurin inhibitor and 8 weeks or longer of consecutive cyclophosphamide use), (4) using time-dependent Cox proportional hazards models to adjust for treatment nonadherence as a time-varying covariate, (5) using a frailty model including a participant-level random effect to account for clustering (of participants included in both treatment arms), (6) performing propensity score weighting using inverse probability weighting with stabilized and truncated (1st/99th percentiles) weights, and (7) defining hypertension as 2 consecutive visits with stage 1 positive hypertensive blood pressure.

Results

Baseline Characteristics

Among 613 INSIGHT participants diagnosed with nephrotic syndrome from 1996 through 2019, 35 were excluded for steroid-resistant disease. Of the remaining 578 participants (median age at diagnosis, 3.7 [IQR, 2.8-6.0] years; 371 male [64%] and 207 female [36%]), 252 initiated cyclophosphamide, 131 initiated calcineurin inhibitors, and 87 initiated both medications sequentially during follow-up. Cyclophosphamide initiation occurred throughout the study period, whereas cyclosporine was the only calcineurin inhibitor used before 2004 and tacrolimus was most common after 2004 (eFigure 1 in Supplement 1). Median (quartile 1 [Q1]; quartile 3 [Q3]) time from nephrotic syndrome diagnosis to medication initiation was 1.0 (0.6; 1.8) years for cyclophosphamide and 1.6 (0.8; 3.6) years for calcineurin inhibitors. Median (Q1; Q3) treatment duration was 9 (8; 12) weeks for cyclophosphamide and 2.3 (1.7; 3.9) years for calcineurin inhibitors. After propensity score weighting, population overlap was achieved (eFigure 2 in Supplement 1) and all baseline covariates were well balanced (Table 1; eFigure 3 in Supplement 1).

Table 1. Baseline Characteristics of Children Diagnosed With Nephrotic Syndrome Who Initiated Cyclophosphamide or a Calcineurin Inhibitor From 1996 to 2019, Before and After Propensity Score Weighting.

Baseline characteristic	Weighting
	Before, No. (%)			After (ie, emulated trial), %^a
	CYP (n = 252)	CNI (n = 131)	Standardized difference^b	CYP	CNI	Standardized difference^b
Demographic variables
Age at medication initiation, y, mean (SD)	6.0 (3.3)	7.2 (3.7)	0.36	7.0 (4.1)	6.9 (4.0)	0.03
Sex
Female	89 (35.3)	48 (36.6)	0.03	36.50	34.10	0.02
Male	163 (64.7)	83 (63.4)	0.03	36.50	34.10	0.02
Ethnicity^c
East or Southeast Asian	18 (7.1)	13 (9.9)	0.19	14.30	14.60	0.02
European	65 (25.8)	37 (28.2)		28.70	29.00
South Asian	94 (37.3)	38 (29.0)		27.60	25.80
Mixed, other,^d or unknown	75 (29.8)	43 (32.8)		29.40	30.70
Immigrant	40 (15.9)	21 (16.0)	0.004	16.20	17.10	0.01
Geographic area^e
Metropolitan Toronto	78 (31.0)	44 (33.6)	0.21	33.60	35.50	0.02
Central or Southwestern Ontario	152 (60.3)	72 (55.0)		53.60	51.50
Eastern Ontario	9 (3.6)	3 (2.3)		4.60	4.60
Northern Ontario	8 (3.2)	5 (3.8)		3.40	3.70
Rural residence	12 (4.8)	8 (6.1)	0.06	8.20	8.40	<0.01
Parental education^e
Less than high school	20 (7.9)	11 (8.4)	0.1	6.40	6.60	0.02
High school	106 (42.1)	53 (40.5)		36.10	34.00
Undergraduate	72 (28.6)	34 (26.0)		30.60	30.40
Graduate	46 (18.3)	28 (21.4)		22.10	24.20
Clinical variables
Community hospital site	61 (24.2)	16 (12.2)	0.32	13.50	12.00	0.01
Body mass index^f z score, mean (SD)^g	1.38 (1.42)	1.64 (1.27)	0.19	1.64 (1.56)	1.66 (1.40)	0.02
Hospitalization at nephrotic syndrome diagnosis	102 (40.5)	51 (38.9)	0.20	40.50	38.50	0.02
Nephrotic syndrome classification prior to medication initiation
Steroid dependent	97 (38.5)	80 (61.1)	0.46	51.80	52.10	0.01
Frequently relapsing	86 (34.1)	28 (21.4)		19.10	18.40
Infrequently relapsing	69 (27.4)	23 (17.6)		29.10	29.50
Days from nephrotic syndrome diagnosis to 1st relapse, mean (SD)	201 (206)	235 (448)	0.1	209.7 (253.2)	205.3 (330.3)	0.02
Days from nephrotic syndrome diagnosis to medication initiation, mean (SD)	588.3 (671.6)	973.4 (983.7)	0.46	809.9 (1075.0)	776.0 (922.1)	0.03
No. of relapses prior to medication initiation, mean (SD)	4.5 (3.8)	6.8 (5.2)	0.51	5.2 (4.9)	5.1 (5.6)	0.02
Blood pressure classification prior to medication initiation^e
Normal blood pressure	91 (36.1)	49 (37.4)	0.29	28.30	29.00	0.01
Elevated blood pressure	41 (16.3)	22 (16.8)		16.00	15.00
Stage 1 hypertension	67 (26.6)	28 (21.4)		23.50	23.60
Stage 2 hypertension	29 (11.5)	9 (6.9)		11.30	11.50
Estimated GFR prior to medication initiation, mean (SD), mL/min/1.73 m²^h	104.4 (17.2)	106.8 (18.2)	0.13	107.3 (16.4)	107.4 (20.0)	0.01
History of atopic condition	113 (44.8)	58 (44.3)	0.01	41.60	41.50	<0.01
Family history of hypertension	12 (4.8)	7 (5.3)	0.03	7.30	7.80	0.01
Family history of diabetes	10 (4.0)	7 (5.3)	0.07	4.50	5.00	0.01
Family history of kidney disease	3 (1.2)	1 (0.8)	0.04	2.70	2.60	<0.01
Medication variables
Medication initiation era
Before 2006	68 (27.0)	21 (16.0)	0.32	16.40	16.90	0.02
2006-2011	50 (19.8)	37 (28.2)		24.20	25.40
2011-2015	72 (28.6)	33 (25.2)		28.80	29
After 2015	62 (24.6)	40 (30.5)		30.70	28.8
Initial prednisone treatment duration after diagnosis^e
<8 wk	21 (8.3)	10 (7.6)	0.21	7.00	6.50	0.03
8-16 wk	129 (51.2)	57 (43.5)		43.40	46.00
>16 wk	70 (27.8)	38 (29.0)		28.20	26.00
Maintenance prednisone use within 1 y prior to medication initiation	21 (8.3)	26 (19.8)	0.4	8.00	8.20	<0.01
Prior nonsteroid immunosuppressive medication use	4 (1.6)	80 (61.1)	1.67	11.60	12.70	0.01
Antihypertensive medication use	11 (4.4)	7 (5.3)	0.14	5.90	6.20	<0.01

Open in a new tab

Abbreviations: CYP, cyclophosphamide; CNI, calcineurin inhibitor; GFR, glomerular filtration rate.

^{^a}

After propensity score weighting, baseline characteristics are presented as percentages and mean (SD). The treatment arms before and after propensity score weighting contain the same individuals. The exact sample size for each treatment arm after weighting is not reported, since these are pseudo-populations (consists of parts of whole individuals based on their propensity score weights).

^{^b}

An absolute standardized mean difference ≥0.1 is considered a substantial difference between groups.^40,41

^{^c}

Ethnicity was self-reported.

^{^d}

Includes Caribbean, African, Central and South American, Indigenous, Middle Eastern, and West Indian.

^{^e}

Missing categorical data were categorized as missing for propensity score analysis. Missing baseline data occurred for postal code (12 children [3.1%]), parental education (13 children [3.4%]), blood pressure (47 children [12.3%]), initial prednisone duration (58 children [15.1%]), and antihypertensive medication use (72 children [18.8%]).

^{^f}

Calculated as weight in kilograms divided by height in meters squared.

^{^g}

Baseline body mass index z-score was missing in 36 children (12.0%). Single imputation of 1.47 (mean value) was performed for propensity score analysis.

^{^h}

Baseline eGFR was missing in 139 children (36.3%). Single imputation of eGFR 105 mL/min/1.73 m² (mean value) was performed for propensity score analysis.

Primary and Secondary Outcomes in the Unweighted Population

Median (Q1; Q3) follow-up after medication initiation was 5.5 (2.5; 9.2) years. Unweighted estimates of all outcomes are presented in Table 2. Relapses occurred in 180 cyclophosphamide-treated children (71%) vs 115 calcineurin inhibitor-treated children (88%) throughout follow-up. The unadjusted hazard ratio (HR) for time to relapse was 1.24 (95% CI, 0.98-1.57) among those treated with calcineurin inhibitors vs cyclophosphamide (Figure 1A). Incident chronic kidney disease was rare after either treatment (5% or less). No patients died during study follow-up, so death was not considered a competing risk.

Table 2. Primary and Secondary Outcomes of Children Diagnosed With Nephrotic Syndrome Who Initiated Cyclophosphamide or a Calcineurin Inhibitor From 1996 to 2019.

Outcome	Cohorts
	Unweighted, No. (%)		Propensity score weighted, weighted effect estimate (comparing CNI vs CYP arms) (95% CI)^a^,^b
	CYP arm (n = 252)	CNI arm (n = 131)
Primary outcome
Time to relapse, y, median (95% CI)	0.8 (0.6-1.1)	1.0 (0.6-1.5)	HR, 1.25 (0.84-1.87)
Control outcomes
Negative control—inhaler use	17 (6.8)	10 (7.6)	HR, 1.26 (0.43-3.65)
Negative—vitamin or supplement use	100 (39.7)	61 (46.6)	HR, 1.06 (0.65-1.72)
Positive control—serum creatinine measurement	180 (71.4)	128 (97.7)	HR, 1.76 (1.23-2.52)
Secondary outcomes
Relapse count per child, median (IQR)
By 1 y	1 (0-2)	1 (0-1)	RR, 1.02 (0.68-1.52)
By 2 y	2 (0-3)	1 (0-3)	RR, 0.91 (0.61-1.35)
By 5 y	3 (0-7)	4 (2-8)	RR, 1.04 (0.69-1.57)
Relapse occurrence
By 1 y^c	132/235 (56.2)	56/119 (47.1)	OR, 0.88 (0.42-1.85)
By 2 y^c	150/212 (70.8)	68/109 (62.4)	OR, 0.55 (0.25-1.23)
By 5 y^c	115/130 (88.5)	72/75 (96.0)	OR, 1.61 (0.36-7.34)
Throughout follow-up	180 (71.4)	115 (87.8)	NA
Safety outcomes
Incident chronic kidney disease	6 (2.4)	7 (5.3)	HR, 1.29 (0.25-6.68)
Annualized eGFR slope (eGFR change from medication initiation to last measurement), mean (SD)	3.4 (11.8)	2.9 (14.0)	β estimate: 2.10 (−3.27 to 7.46)
Any subsequent steroid-sparing medication use	111 (44.1)	63 (48.1)	HR, 0.73 (0.45-1.18)
Cyclophosphamide	0	3 (2.3)	NA
Calcineurin inhibitor	75 (29.8)	3 (2.3)	NA
Mycophenolate mofetil	29 (11.5)	38 (29.0)	NA
Rituximab	1 (0.4)	19 (14.5)	NA
Levamisole	6 (2.4)	0	NA
Hospitalization	106 (42.1)	89 (67.9)	HR, 1.83 (1.14-2.92)
Intravenous albumin	71 (28.2)	68 (51.9)	HR, 2.81 (1.65-4.81)
Infection	168 (66.7)	104 (79.4)	HR, 0.91 (0.62-1.33)
Hypertensive blood pressure or incident anti-hypertensive medication use	111 (44.1)	72 (55.0)	HR, 1.08 (0.68-1.72)
Incident anti-hypertensive medication use	31 (12.3)	31 (23.7)	HR, 1.82 (0.83-4.01)
PedsQL total score postmedication, mean (SD)
Any time during follow-up	86.2 (11.6)	84.2 (13.2)	NA
By 1 y	88.4 (11.4)	86.2 (13.4)	NA
By 2 y	88.3 (10.9)	85.9 (13.1)	NA
Change in PedsQL score after medication initiation, mean (SD)	1.1 (10.5)	1.7 (10.7)	β estimate: 2.90 (−4.05 to 9.85)

Open in a new tab

Abbreviations: CNI, calcineurin inhibitor; CYP, cyclophosphamide; eGFR, estimated glomerular filtration rate; HR, hazard ratio; NA, not applicable; PedsQL, pediatric quality of life; RR, rate ratio; OR, odds ratio.

^{^a}

Comparison between CNI arm and CYP arm (reference group). For example, an HR of 1.28 indicates a 28% higher hazard rate for relapse among children treated with calcineurin inhibitors, compared with those treated with cyclophosphamide.

^{^b}

95% CIs were calculated using robust sandwich to type variance estimation.

^{^c}

Among participants with complete follow-up to the respective time point.

Figure 1. — Time to relapse was analyzed by weighted Kaplan-Meier method. Shaded areas represent the 95% CIs (using robust variance estimation) around each point estimate.

Relapses in the Weighted Emulated Trial Population

There was no significant difference in time to relapse between children treated with calcineurin inhibitors vs cyclophosphamide after propensity score weighting. The weighted HR for time to relapse was 1.25 (95% CI, 0.84-1.87; P = .30) among those treated with calcineurin inhibitors vs cyclophosphamide (Table 2, Figure 1B). There were also no significant differences in relapse count or relapse occurrence by year 1, year 2, or year 5. In the weighted cohorts, 62%, 64%, and 85% of children relapsed by year 1, year 2, and throughout follow-up after calcineurin inhibitor initiation, compared with 57%, 68%, and 73% after cyclophosphamide. In stratified analyses, there was no evidence of treatment effect heterogeneity by sex, age, ethnicity, nephrotic syndrome classification, or clinical center type (Figure 2).

Figure 2. — Effect modification was evaluated using stratified analysis with subgroup-balancing propensity score methods. The forest plot presents the effect estimate (weighted hazard ratio [HR] for time to relapse) and 95% CIs for each subgroup, comparing CNI vs CYP treatment on a logarithmic scale. An HR more than 1 indicates an increased hazard rate for relapse among children treated with CNI compared with those treated with CYP. The relapse percentage columns on the right indicate the proportion of children in each subgroup who experience a subsequent relapse at any time during follow-up after initiation of either medication. An age cutoff of younger than 8 or 8 years and older was selected for comparison to prior studies.^46,47,48 FR/SDNS indicates frequently relapsing or steroid dependent nephrotic syndrome; IRNS, infrequently relapsing nephrotic syndrome.

Secondary Outcomes in the Weighted Emulated Trial Population

There were no significant differences in incident chronic kidney disease, eGFR slope, or subsequent immunosuppressive medication use between children treated with cyclophosphamide vs calcineurin inhibitors (Table 2). Children initially treated with cyclophosphamide typically received calcineurin inhibitors, whereas children treated with calcineurin inhibitors typically received mycophenolate or rituximab. Calcineurin inhibitor treatment was associated with a higher risk of hospitalization (HR, 1.83; 95% CI, 1.14-2.92; eFigure 4 in Supplement 1) and intravenous albumin use (HR, 2.81; 95% CI, 1.65-4.81; eFigure 5 in Supplement 1) than cyclophosphamide. Hospitalizations were most common during the first year of calcineurin inhibitor treatment and 138 of 195 hospitalized children relapsed before hospitalization (71% ). There were also no significant differences in change in quality-of-life scores or adverse events, including infection, hypertension diagnosis, or antihypertensive treatment between medications.

Residual Confounding

E-values and control outcomes were used to evaluate for residual confounding. The E-values for the time to relapse point estimate and confidence interval were 1.61 and 1.00, respectively. Thus, 1.61 is the minimum strength of association (on an HR scale) that a set of unmeasured confounders must have with both the treatment and outcome in order to nullify the observed treatment-outcome association (ie, HR 1.25 for time-to-relapse between treatments).⁴³ The E-value for the confidence interval is 1, since no significant treatment-outcome association was observed. Treatment assignment was not associated with negative control outcomes inhaler use (weighted HR, 1.26; 95% CI, 0.43-3.65) or vitamin/supplement use (weighted HR, 1.06; 95% CI, 0.65-1.72; eFigure 6 in Supplement 1). Calcineurin inhibitor treatment was significantly associated with the positive control outcome: serum creatinine measurement (weighted HR, 1.76; 95% CI, 1.23-2.52).

Sensitivity Analyses

The robustness of the results were tested to various assumptions, including participant clustering, treatment nonadherence, and propensity score specification (eTable 2, eFigures 7 through 10 in Supplement 1). There were no significant differences in time to relapse after cyclophosphamide vs calcineurin inhibitor initiation in any sensitivity analysis. Specifically, there was no difference when restricting to the first use of either medication (weighted HR, 1.31; 95% CI, 0.82-2.11) or tacrolimus vs cyclophosphamide (weighted HR, 1.29; 95% CI, 0.83-2.02). There was also no difference in time to hypertension diagnosis or antihypertensive medication between arms, when defining hypertension as 2 consecutive visits with stage 1 positive hypertensive blood pressure (eFigure 11 in Supplement 1).

Discussion

In this target trial emulation analysis, there was no significant difference in the effectiveness of cyclophosphamide vs calcineurin inhibitors for preventing childhood nephrotic syndrome relapses. Although most children experienced subsequent relapses and half received additional nonsteroid immunosuppressive medications, the risk of developing chronic kidney disease was low (5% or less) after either treatment. Therefore, the choice between cyclophosphamide or calcineurin inhibitor treatment should involve shared decision-making between physicians, patients, and their family. Our findings have important global health implications, since calcineurin inhibitor use is associated with higher patient and health care system burden related to medication cost and monitoring, without clear evidence of superiority.²⁰

There is limited comparative effectiveness data for nonsteroid immunosuppressants in childhood nephrotic syndrome. Alkylating agents and calcineurin inhibitors are the most common medications used worldwide, but only 2 small RCTs have compared them.^15,49 Among 40 children with steroid-dependent nephrotic syndrome, relapses were more common by 2 years after 6 months of cyclosporine vs 6 weeks of chlorambucil treatment (95% vs 55%).¹⁵ Cyclosporine treatment also resulted in more complications, including hypertension, acute kidney injury, and cosmetic adverse effects. Among 73 adults and children with steroid-sensitive nephrotic syndrome, relapses by 2 years were more common after 1 year of cyclosporine vs 8 weeks of cyclophosphamide (75% vs 37%).⁴⁹ There are limited data on tacrolimus, which is associated with fewer complications and may prolong remission.^50,51,52,53 Using target trial emulation, we found no significant difference in time to relapse after cyclophosphamide or calcineurin inhibitor treatment. Findings were consistent and robust to multiple sensitivity analyses, including a direct comparison of cyclophosphamide vs tacrolimus. Relapse frequency decreased substantially after initiating either treatment, yet less than 30% of children experience long-term remission.

Children treated with calcineurin inhibitors were at higher risk of hospitalization and intravenous albumin administration. Hospitalizations typically occurred within 1 year of calcineurin inhibitor initiation, after children had relapsed. Children who relapse while taking calcineurin inhibitors may be hospitalized and administered intravenous albumin to prevent or treat intravascular volume depletion and acute kidney injury. Nephrotoxic medications, including calcineurin inhibitors, are associated with acute kidney injury during childhood nephrotic syndrome hospitalizations.⁵⁴ Calcineurin inhibitor treatment may cause other treatment-related complications (eg, severe hypertension or laboratory abnormalities) warranting hospitalization.

Immunosuppressive treatment response varies in nephrotic syndrome. Older age and less steroid dependence are associated with prolonged remission after cyclophosphamide treatment.^46,47 Younger age, lower prior relapse rate, and less steroid dependence are associated with prolonged remission after calcineurin inhibitor treatment.^13,55 We did not observe treatment-effect heterogeneity by age, sex, ethnicity, disease classification, or clinical center in our study, despite the larger sample size than prior reports. Other factors, including genetic polymorphisms and auto-antibodies, may explain heterogeneity in immunosuppression response.^56,57

Our study found that cyclophosphamide and calcineurin inhibitors were similarly effective at preventing nephrotic syndrome relapses. Yet, calcineurin inhibitors are considerably more expensive. For a 20-kg child, the medication cost for 2 years of tacrolimus is CAD $1900 (US $1319) and cyclosporine is CAD $2250 (US $1561) vs CAD $27 (US $19) for 8 weeks of cyclophosphamide.⁵⁸ Furthermore, calcineurin inhibitor treatment leads to additional hospitalizations and health care system cost, patient burden, and caregiver productivity loss due to laboratory surveillance (at least every 3 months)¹⁰ and protocolized kidney biopsy procedures (in some centers). Thus, calcineurin inhibitors are inaccessible to children in many low- to middle-income countries.²⁰

Despite this, calcineurin inhibitors are increasingly used to treat nephrotic syndrome, based on concerns about long-term infertility after cyclophosphamide.^21,22,23,24 However, this perceived risk may be exaggerated.^23,59 Among childhood cancer survivors, only cumulative cyclophosphamide equivalent doses (CED) of 11.3 g/m² or more are associated with a lower likelihood of pregnancy.⁶⁰ This is more than double the dose typically used in childhood nephrotic syndrome (168 mg/kg equals approximately CED 4.5 g/m²). Among males, a CED of 7.5 g/m² or more is associated with gonadal toxicity.^21,61 Lower cyclophosphamide doses are also associated with oligospermia, although this is often reversible and has not been associated with permanent infertility.^61,62,63,64 Lastly, using a prediction model for ovarian failure developed among 6761 female childhood cancer survivors,⁶⁵ the estimated risk of ovarian failure with typical nephrotic syndrome cyclophosphamide dosing (168 mg/kg) is 2.4%.⁶⁵ Thus, concerns about cyclophosphamide-related infertility in childhood nephrotic syndrome may be overstated and should be balanced against its potential benefits, including lower cost, longer medication-free remission, and global accessibility. Additional long-term observational data are needed to determine the risk of infertility in this population.

Strengths and Limitations

Our study has several limitations. Based on local treatment protocols, most children received cyclophosphamide before calcineurin inhibitors. Participants who received both medications were included in both arms, with separate propensity scores calculated using updated covariates. We also performed sensitivity analyses restricting to first medication use and accounting for clustered observations, which did not change our results. Our initial tacrolimus trough-level target of 3 to 5 ng/mL is consistent with current guidelines¹⁰ but limits the generalizability of our findings for centers that target higher initial tacrolimus trough levels (ie, 5 to 10 ng/mL). Specific outcomes (hospitalization, infection, inhaler, and vitamin/supplement use) were only assessed at annual study visits and the visit date was used as an event proxy. Relapses could be missed if home urine monitoring was not performed and spontaneous remission occurred or if a relapse treated at another clinical center was not documented. However, both scenarios are uncommon in our nursing-based practice. Missing baseline covariates were addressed using mean imputation and missing categories to estimate propensity scores. Subgroup analyses were underpowered to detect meaningful differences in relapses. Based on sample size, steroid-dependent and frequently relapsing disease categories were combined to facilitate subgroup balancing propensity score methods. The proportion of children classified as infrequently relapsing (24%) was higher than expected. Among them, the median relapse rate premedication was 2.3 relapses per year (Q1; Q3, 1.5; 3.2) and 36% of them would be reclassified as frequently relapsing using the recent International Pediatric Nephrology Association definition.¹⁰ The indication for steroid-sparing medication use in these children was determined by the treating physician, but was likely based on relapse frequency, disease complications, or steroid-related toxicity.

This study also has multiple strengths. We included a large childhood nephrotic syndrome cohort initiating cyclophosphamide or calcineurin inhibitors that was adequately powered to detect a clinically important difference in time to relapse between treatments. Based on the available sample size (N = 383; power = 0.8 and α = .05), this study could detect a 39% relative difference in time-to-relapse (HR, 1.39; approximately a 10-week difference in time to relapse) between treatments. Enrollment at multiple academic and community sites from a diverse socioeconomic area increases generalizability. Median follow-up duration was more than 5 years in both groups, facilitating analysis of long-term outcomes not possible in previous studies. Target trial emulation using propensity score methods allowed us to prevent common observational research biases, balance many measured confounders, and verify covariate balance. Overlap weighting increases generalizability to individuals at clinical equipoise. This relatively upweighted children who had not received prior nonsteroid immunosuppressive medications.

Conclusions

In this study, cyclophosphamide and calcineurin inhibitors were similarly effective for preventing childhood nephrotic syndrome relapses. These findings can inform therapeutic decision-making and patient counseling. The safety, cost, and burden of the 2 medications should be considered when selecting between them, particularly in resource-limited practice settings. Lastly, target trial emulation is a valuable research tool which can be used to compare the effectiveness of different treatments in common childhood diseases, when RCT data are not available or feasible to conduct.

Supplement 1.

eTable 1. Specification and emulation of a target pragmatic clinical trial comparing cyclophosphamide versus calcineurin inhibitor treatment in childhood nephrotic syndrome

eTable 2. Sensitivity analyses for the primary outcome (time-to-relapse) among children diagnosed with nephrotic syndrome that initiated cyclophosphamide or a calcineurin inhibitor from 1996 to 2019

eFigure 1. Trends in steroid-sparing medication selection over the course of the INSIGHT study period

eFigure 2. Propensity score distributions before (left) and after (right) overlap weighting procedure, among cyclophosphamide and calcineurin inhibitor arms

eFigure 3. Covariate balance before and after propensity score overlap weighting

eFigure 4. Cumulative probability of hospitalization after cyclophosphamide vs. calcineurin inhibitor initiation

eFigure 5. Cumulative probability of intravenous albumin administration after cyclophosphamide vs. calcineurin inhibitor initiation

eFigure 6. Cumulative probability of a) inhaler use (negative control), b) vitamin or supplement use (negative control), and c) serum creatinine measurement (positive control)

eFigure 7. Time-to-relapse after the first use of either cyclophosphamide vs. calcineurin inhibitor a) prior to propensity score weighting and b) after propensity score weighting

eFigure 8. Time-to-relapse after cyclophosphamide vs. tacrolimus initiation after Jan 1, 2004 a) prior to propensity score weighting and b) after propensity score weighting

eFigure 9. Time-to-relapse after “per-protocol” use of cyclophosphamide vs. calcineurin inhibitor a) prior to propensity score weighting and b) after propensity score weighting

eFigure 10. Time-to-relapse after cyclophosphamide vs. calcineurin inhibitor initiation after inverse probability of treatment weighting (stabilized and truncated weights)

eFigure 11. Cumulative probability of hypertension diagnosis (two consecutive clinic visits with blood pressure measurements in the stage 1-2 hypertension range) or anti-hypertensive medication initiation after cyclophosphamide vs. calcineurin inhibitor initiation

eAppendix 1. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist

jamapediatr-e245286-s001.pdf^{(809.2KB, pdf)}

Supplement 2.

Data sharing statement

jamapediatr-e245286-s002.pdf^{(13.2KB, pdf)}

References

1.Noone DG, Iijima K, Parekh R. Idiopathic nephrotic syndrome in children. Lancet. 2018;392(10141):61-74. doi: 10.1016/S0140-6736(18)30536-1 [DOI] [PubMed] [Google Scholar]
2.Veltkamp F, Rensma LR, Bouts AHM; LEARNS consortium . Incidence and relapse of idiopathic nephrotic syndrome: meta-analysis. Pediatrics. 2021;148(1):e2020029249. doi: 10.1542/peds.2020-029249 [DOI] [PubMed] [Google Scholar]
3.Sinha A, Hari P, Sharma PK, et al. Disease course in steroid sensitive nephrotic syndrome. Indian Pediatr. 2012;49(11):881-887. doi: 10.1007/s13312-012-0220-4 [DOI] [PubMed] [Google Scholar]
4.Özlü SG, Demircin G, Tökmeci N, et al. Long-term prognosis of idiopathic nephrotic syndrome in children. Ren Fail. 2015;37(4):672-677. doi: 10.3109/0886022X.2015.1010940 [DOI] [PubMed] [Google Scholar]
5.Carter SA, Mistry S, Fitzpatrick J, et al. Prediction of short- and long-term outcomes in childhood nephrotic syndrome. Kidney Int Rep. 2019;5(4):426-434. doi: 10.1016/j.ekir.2019.12.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Khullar S, Banh T, Vasilevska-Ristovska J, et al. Impact of steroids and steroid-sparing agents on quality of life in children with nephrotic syndrome. Pediatr Nephrol. 2021;36(1):93-102. doi: 10.1007/s00467-020-04684-3 [DOI] [PubMed] [Google Scholar]
7.Primary nephrotic syndrome in children: clinical significance of histopathologic variants of minimal change and of diffuse mesangial hypercellularity. A report of the International Study of Kidney Disease in Children. Kidney Int. 1981;20(6):765-771. doi: 10.1038/ki.1981.209 [DOI] [PubMed] [Google Scholar]
8.Banh THM, Hussain-Shamsy N, Patel V, et al. Ethnic differences in incidence and outcomes of childhood nephrotic syndrome. Clin J Am Soc Nephrol. 2016;11(10):1760-1768. doi: 10.2215/CJN.00380116 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Rovin BH, Adler SG, Barratt J, et al. ; Kidney Disease: Improving Global Outcomes (KDIGO) Glomerular Diseases Work Group . KDIGO 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int. 2021;100(4S):S1-S276. doi: 10.1016/j.kint.2021.05.021 [DOI] [PubMed] [Google Scholar]
10.Trautmann A, Boyer O, Hodson E, et al. ; International Pediatric Nephrology Association . IPNA clinical practice recommendations for the diagnosis and management of children with steroid-sensitive nephrotic syndrome. Pediatr Nephrol. 2023;38(3):877-919. doi: 10.1007/s00467-022-05739-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Larkins NG, Liu ID, Willis NS, Craig JC, Hodson EM. Non-corticosteroid immunosuppressive medications for steroid-sensitive nephrotic syndrome in children. Cochrane Database Syst Rev. 2020;4(4):CD002290. doi: 10.1002/14651858.CD002290.pub5 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Tan L, Li S, Yang H, Zou Q, Wan J, Li Q. Efficacy and acceptability of immunosuppressive agents for pediatric frequently-relapsing and steroid-dependent nephrotic syndrome: a network meta-analysis of randomized controlled trials. Medicine (Baltimore). 2019;98(22):e15927. doi: 10.1097/MD.0000000000015927 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Ishikura K, Yoshikawa N, Nakazato H, et al. ; Japanese Study Group of Renal Disease in Children . Two-year follow-up of a prospective clinical trial of cyclosporine for frequently relapsing nephrotic syndrome in children. Clin J Am Soc Nephrol. 2012;7(10):1576-1583. doi: 10.2215/CJN.00110112 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Kitano Y, Yoshikawa N, Tanaka R, Nakamura H, Ninomiya M, Ito H. Ciclosporin treatment in children with steroid-dependent nephrotic syndrome. Pediatr Nephrol. 1990;4(5):474-477. doi: 10.1007/BF00869823 [DOI] [PubMed] [Google Scholar]
15.Niaudet P; The French Society of Paediatric Nephrology . Comparison of cyclosporin and chlorambucil in the treatment of steroid-dependent idiopathic nephrotic syndrome: a multicentre randomized controlled trial. Pediatr Nephrol. 1992;6(1):1-3. doi: 10.1007/BF00856817 [DOI] [PubMed] [Google Scholar]
16.Tanaka R, Yoshikawa N, Kitano Y, Ito H, Nakamura H. Long-term ciclosporin treatment in children with steroid-dependent nephrotic syndrome. Pediatr Nephrol. 1993;7(3):249-252. doi: 10.1007/BF00853209 [DOI] [PubMed] [Google Scholar]
17.Samuel S, Morgan CJ, Bitzan M, et al. Substantial practice variation exists in the management of childhood nephrotic syndrome. Pediatr Nephrol. 2013;28(12):2289-2298. doi: 10.1007/s00467-013-2546-0 [DOI] [PubMed] [Google Scholar]
18.Schijvens AM, van der Weerd L, van Wijk JAE, et al. Practice variations in the management of childhood nephrotic syndrome in the Netherlands. Eur J Pediatr. 2021;180(6):1885-1894. doi: 10.1007/s00431-021-03958-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Sinha A, Bagga A. Clinical practice guidelines for nephrotic syndrome: consensus is emerging. Pediatr Nephrol. 2022;37(12):2975-2984. doi: 10.1007/s00467-022-05639-6 [DOI] [PubMed] [Google Scholar]
20.Ademola AD, Asinobi AO, Alao MA, Olowu WA. Childhood nephrotic syndrome in Africa: epidemiology, treatment trends, and outcomes. Semin Nephrol. 2022;42(5):151311. doi: 10.1016/j.semnephrol.2023.151311 [DOI] [PubMed] [Google Scholar]
21.Kenney LB, Laufer MR, Grant FD, Grier H, Diller L. High risk of infertility and long term gonadal damage in males treated with high dose cyclophosphamide for sarcoma during childhood. Cancer. 2001;91(3):613-621. doi: [DOI] [PubMed] [Google Scholar]
22.Poorvu PD, Frazier AL, Feraco AM, et al. Cancer Treatment-related infertility: a critical review of the evidence. J Natl Cancer Inst Cancer Spectr. 2019;3(1):pkz008. doi: 10.1093/jncics/pkz008 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Gajjar R, Miller SD, Meyers KE, Ginsberg JP. Fertility preservation in patients receiving cyclophosphamide therapy for renal disease. Pediatr Nephrol. 2015;30(7):1099-1106. doi: 10.1007/s00467-014-2897-1 [DOI] [PubMed] [Google Scholar]
24.van den Brand JA, van Dijk PR, Hofstra JM, Wetzels JF. Cancer risk after cyclophosphamide treatment in idiopathic membranous nephropathy. Clin J Am Soc Nephrol. 2014;9(6):1066-1073. doi: 10.2215/CJN.08880813 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Kari JA, Alhasan KA, Albanna AS, et al. Rituximab versus cyclophosphamide as first steroid-sparing agent in childhood frequently relapsing and steroid-dependent nephrotic syndrome. Pediatr Nephrol. 2020;35(8):1445-1453. doi: 10.1007/s00467-020-04570-y [DOI] [PubMed] [Google Scholar]
26.Sandhu J, Bhat D, Dhooria GS, et al. Oral cyclophosphamide therapy in 100 children with steroid-sensitive nephrotic syndrome: experience from a developing country. Pediatr Nephrol. 2021;36(9):2759-2767. doi: 10.1007/s00467-021-05052-5 [DOI] [PubMed] [Google Scholar]
27.Modi ZJ, Zhai Y, Yee J, et al. ; NEPTUNE investigators . Pediatric contributions and lessons learned from the NEPTUNE cohort study. Pediatr Nephrol. 2024;39(9):2555-2568. doi: 10.1007/s00467-023-06256-7 [DOI] [PubMed] [Google Scholar]
28.Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available. Am J Epidemiol. 2016;183(8):758-764. doi: 10.1093/aje/kwv254 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Hernán MA, Sauer BC, Hernández-Díaz S, Platt R, Shrier I. Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses. J Clin Epidemiol. 2016;79:70-75. doi: 10.1016/j.jclinepi.2016.04.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Kutcher SA, Brophy JM, Banack HR, Kaufman JS, Samuel M. Emulating a randomised controlled trial with observational data: an introduction to the target trial framework. Can J Cardiol. 2021;37(9):1365-1377. doi: 10.1016/j.cjca.2021.05.012 [DOI] [PubMed] [Google Scholar]
31.Hussain N, Zello JA, Vasilevska-Ristovska J, et al. The rationale and design of Insight into Nephrotic Syndrome: Investigating Genes, Health and Therapeutics (INSIGHT): a prospective cohort study of childhood nephrotic syndrome. BMC Nephrol. 2013;14:25. doi: 10.1186/1471-2369-14-25 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453-1457. doi: 10.1016/S0140-6736(07)61602-X [DOI] [PubMed] [Google Scholar]
33.Pierce CB, Muñoz A, Ng DK, Warady BA, Furth SL, Schwartz GJ. Age- and sex-dependent clinical equations to estimate glomerular filtration rates in children and young adults with chronic kidney disease. Kidney Int. 2021;99(4):948-956. doi: 10.1016/j.kint.2020.10.047 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.KDIGO CKD Work Group . CKD evaluation and management. Accessed December 13, 2024. https://kdigo.org/guidelines/ckd-evaluation-and-management/
35.Flynn JT, Kaelber DC, Baker-Smith CM, et al. ; SUBCOMMITTEE ON SCREENING AND MANAGEMENT OF HIGH BLOOD PRESSURE IN CHILDREN . Clinical practice guideline for screening and management of high blood pressure in children and adolescents. Pediatrics. 2017;140(3):e20171904. doi: 10.1542/peds.2017-1904 [DOI] [PubMed] [Google Scholar]
36.Varni JW, Seid M, Kurtin PS. PedsQL 4.0: reliability and validity of the Pediatric Quality of Life Inventory version 4.0 generic core scales in healthy and patient populations. Med Care. 2001;39(8):800-812. doi: 10.1097/00005650-200108000-00006 [DOI] [PubMed] [Google Scholar]
37.Thomas LE, Li F, Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA. 2020;323(23):2417-2418. doi: 10.1001/jama.2020.7819 [DOI] [PubMed] [Google Scholar]
38.Li F, Thomas LE, Li F. Addressing extreme propensity scores via the overlap weights. Am J Epidemiol. 2019;188(1):250-257. doi: 10.1093/aje/kwy201 [DOI] [PubMed] [Google Scholar]
39.Desai RJ, Franklin JM. Alternative approaches for confounding adjustment in observational studies using weighting based on the propensity score: a primer for practitioners. BMJ. Accessed December 13, 2024. doi: 10.1136/bmj.l5657 [DOI] [PubMed] [Google Scholar]
40.Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput. 2009;38(6):1228-1234. doi: 10.1080/03610910902859574 [DOI] [Google Scholar]
41.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083-3107. doi: 10.1002/sim.3697 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Greifer N. Cobalt: covariate balance tables and plots. Accessed December 13, 2024. https://ngreifer.github.io/cobalt
43.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the e-value. Ann Intern Med. 2017;167(4):268-274. doi: 10.7326/M16-2607 [DOI] [PubMed] [Google Scholar]
44.Greifer N. WeightIt: weighting for covariate balance in observational studies. Accessed December 13, 2024. doi: 10.32614/CRAN.package.WeightIt [DOI]
45.Dong J, Zhang JL, Zeng S, Li F. Subgroup balancing propensity score. Stat Methods Med Res. 2020;29(3):659-676. doi: 10.1177/0962280219870836 [DOI] [PubMed] [Google Scholar]
46.Cammas B, Harambat J, Bertholet-Thomas A, et al. Long-term effects of cyclophosphamide therapy in steroid-dependent or frequently relapsing idiopathic nephrotic syndrome. Nephrol Dial Transplant. 2011;26(1):178-184. doi: 10.1093/ndt/gfq405 [DOI] [PubMed] [Google Scholar]
47.Azib S, Macher MA, Kwon T, et al. Cyclophosphamide in steroid-dependent nephrotic syndrome. Pediatr Nephrol. 2011;26(6):927-932. doi: 10.1007/s00467-011-1830-0 [DOI] [PubMed] [Google Scholar]
48.Zagury A, de Oliveira AL, de Moraes CAP, et al. Long-term follow-up after cyclophosphamide therapy in steroid-dependent nephrotic syndrome. Pediatr Nephrol. 2011;26(6):915-920. doi: 10.1007/s00467-011-1825-x [DOI] [PubMed] [Google Scholar]
49.Ponticelli C, Edefonti A, Ghio L, et al. Cyclosporin versus cyclophosphamide for patients with steroid-dependent and frequently relapsing idiopathic nephrotic syndrome: a multicentre randomized controlled trial. Nephrol Dial Transplant. 1993;8(12):1326-1332. [PubMed] [Google Scholar]
50.Choudhry S, Bagga A, Hari P, Sharma S, Kalaivani M, Dinda A. Efficacy and safety of tacrolimus versus cyclosporine in children with steroid-resistant nephrotic syndrome: a randomized controlled trial. Am J Kidney Dis. 2009;53(5):760-769. doi: 10.1053/j.ajkd.2008.11.033 [DOI] [PubMed] [Google Scholar]
51.Li HY, Zhang X, Zhou T, Zhong Z, Zhong H. Efficacy and safety of cyclosporine a for patients with steroid-resistant nephrotic syndrome: a meta-analysis. BMC Nephrol. 2019;20(1):384. doi: 10.1186/s12882-019-1575-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Wang W, Xia Y, Mao J, et al. Treatment of tacrolimus or cyclosporine A in children with idiopathic nephrotic syndrome. Pediatr Nephrol. 2012;27(11):2073-2079. doi: 10.1007/s00467-012-2228-3 [DOI] [PubMed] [Google Scholar]
53.Sinha MD, MacLeod R, Rigby E, Clark AGB. Treatment of severe steroid-dependent nephrotic syndrome (SDNS) in children with tacrolimus. Nephrol Dial Transplant. 2006;21(7):1848-1854. doi: 10.1093/ndt/gfi274 [DOI] [PubMed] [Google Scholar]
54.Rheault MN, Zhang L, Selewski DT, et al. ; Midwest Pediatric Nephrology Consortium . AKI in children hospitalized with nephrotic syndrome. Clin J Am Soc Nephrol. 2015;10(12):2110-2118. doi: 10.2215/CJN.06620615 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Ishikura K, Yoshikawa N, Hattori S, et al. ; for Japanese Study Group of Renal Disease in Children . Treatment with microemulsified cyclosporine in children with frequently relapsing nephrotic syndrome. Nephrol Dial Transplant. 2010;25(12):3956-3962. doi: 10.1093/ndt/gfq318 [DOI] [PubMed] [Google Scholar]
56.Watts AJB, Keller KH, Lerner G, et al. Discovery of autoantibodies targeting nephrin in minimal change disease supports a novel autoimmune etiology. J Am Soc Nephrol. 2022;33(1):238-252. doi: 10.1681/ASN.2021060794 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Mo X, Chen X, Zeng H, et al. Tacrolimus in the treatment of childhood nephrotic syndrome: machine learning detects novel biomarkers and predicts efficacy. Pharmacotherapy. 2023;43(1):43-52. doi: 10.1002/phar.2749 [DOI] [PubMed] [Google Scholar]
58.Ontario Ministry of Health . Ontario Drug Benefit (ODB) Formulary/Comparative Drug Index (CDI) edition 43. Accessed December 13, 2024. https://www.ontario.ca/document/ontario-drug-benefit-odb-formulary-comparative-drug-index-cdi-and-monthly-formulary-0
59.BC Renal GN Committee . Gonadal Toxicity with Cyclophosphamide. BC Renal Agency; 2019. [Google Scholar]
60.Chow EJ, Stratton KL, Leisenring WM, et al. Pregnancy after chemotherapy in male and female survivors of childhood cancer treated between 1970 and 1999: a report from the Childhood Cancer Survivor Study cohort. Lancet Oncol. 2016;17(5):567-576. doi: 10.1016/S1470-2045(16)00086-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Meistrich ML, Wilson G, Brown BW, da Cunha MF, Lipshultz LI. Impact of cyclophosphamide on long-term reduction in sperm count in men treated with combination chemotherapy for Ewing and soft tissue sarcomas. Cancer. 1992;70(11):2703-2712. doi: [DOI] [PubMed] [Google Scholar]
62.Trompeter RS, Evans PR, Barratt TM. Gonadal function in boys with steroid-responsive nephrotic syndrome treated with cyclophosphamide for short periods. Lancet. 1981;1(8231):1177-1179. doi: 10.1016/S0140-6736(81)92348-5 [DOI] [PubMed] [Google Scholar]
63.Green DM, Liu W, Kutteh WH, et al. Cumulative alkylating agent exposure and semen parameters in adult survivors of childhood cancer: a report from the St Jude Lifetime Cohort Study. Lancet Oncol. 2014;15(11):1215-1223. doi: 10.1016/S1470-2045(14)70408-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Latta K, von Schnakenburg C, Ehrich JH. A meta-analysis of cytotoxic treatment for frequently relapsing nephrotic syndrome in children. Pediatr Nephrol. 2001;16(3):271-282. doi: 10.1007/s004670000523 [DOI] [PubMed] [Google Scholar]
65.Clark RA, Mostoufi-Moab S, Yasui Y, et al. Predicting acute ovarian failure in female survivors of childhood cancer: a cohort study in the Childhood Cancer Survivor Study (CCSS) and the St Jude Lifetime Cohort (SJLIFE). Lancet Oncol. 2020;21(3):436-445. doi: 10.1016/S1470-2045(19)30818-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials