#NephJC Chat
Tuesday June 25, 2019 at 9 pm Eastern Daylight Time
Wednesday June 26, 2019 at 9 pm Indian Standard Time
Wednesday June 26, 2019 at 9 pm British Summer Time
Health Technol Assess. 2019 May;23(26):1-108. doi: 10.3310/hta23260.Paperpile
Sixteen-week versus standard eight-week prednisolone therapy for childhood nephrotic syndrome: the PREDNOS RCT.
Webb NJ, Woolley RL, Lambe T, Frew E, Brettell EA, Barsoum EN, Trompeter RS, Cummins C, Wheatley K, Ives NJ.
PMID: 31156083 Full Text at HTA, and on the BMJ
Introduction
Idiopathic nephrotic syndrome is the most common form of glomerular disease in children. In the 1960-1970s, the International Study of Kidney Disease in Children (ISKDC) study showed that the majority of children who presented with nephrotic syndrome without hypertension, hematuria or reduced kidney function had minimal change disease on kidney biopsy, particularly those who responded to 8 weeks of empiric steroid therapy. This resulted in the current approach to pediatric idiopathic nephrotic syndrome where a trial of high-dose steroid therapy is initiated without a kidney biopsy. However, how much steroids and for how long should they be continued has been a raging debate in the pediatric nephrology community. If you ask 10 pediatric nephrologists the “right” amount of steroids to use to treat idiopathic nephrotic syndrome, you’re likely to get 10 different answers. Not just the dose, but the duration, and the tapering strategy are all practiced as an art more than a science or fixed algorithm.
The 2012 Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guidelines for Glomerulonephritis used the recommendations from the ISKDC study to make the following guideline for treating children aged 1 to 18 years presenting with nephrotic syndrome:
“ a course of corticosteroid therapy for at least 12 weeks, starting with oral prednisone at 60 mg/m2 daily or 2 mg/kg/d (max of 60 mg/d) for 4 to 6 weeks, followed by alternate-day oral prednisolone at 40 mg/m2 or 1.5 mg/kg (max 40 mg) on alternate days continued for 2 to 5 months with a tapering dose.”
Several randomized controlled trials (RCTs) have looked at duration of steroid therapy and its affect on the risk of relapse and the development of frequently relapsing nephrotic syndrome. A Cochrane systematic review on corticosteroid therapy for nephrotic syndrome in children (last updated in 2015) examined randomized controlled trials comparing different treatment durations and dosing strategies. They found differences in treatment effects between older studies with a high risk of bias, and more recent studies with low risk of bias.
In the older, more biased studies, a longer duration of corticosteroid therapy (>3 months) appeared to reduce the risk of relapse and the risk of developing frequently relapsing nephrotic syndrome.
In more recent studies, with lower risk of bias, there was no significant difference between standard (2-3 months) of steroid treatment compared to longer courses of steroids.
Two newer RCTs published in 2015, after the Cochrane systematic review, examined the duration of steroid therapy and risk of relapse. Sinha et al compared the relapse rate in patients receiving 3 months of corticosteroid therapy compared to 6 months of therapy in 5 academic centers in North India. They found no statistically significant difference in relapse rate, frequent relapses or adverse effects due to steroids. Yoshikawa et al performed a multicenter, open-labelled RCT comparing 2 months of prednisolone therapy with 6 months of therapy in 255 children at 90 hospitals in Japan. Similarly, there was no difference in the clinical outcomes including time to first relapse and rate of frequently relapsing nephrotic syndrome.
The Study
Methods
This study was a multi-center, double-blind, placebo-controlled, randomized trial involving 125 general hospitals and tertiary pediatric nephrology centers in the United Kingdom National Health Service. To minimize drop-out due to steroid resistance, which was a problem in some previous studies, all potential participants were treated initially with 60 mg/m2 prednisolone daily for 4 weeks and were recruited and randomized only if the children were corticosteroid sensitive, developed resolution of proteinuria. This typically occurred between 14 and 21 days after starting treatment. In addition, during randomization, a minimization algorithm was used to ensure balanced allocation by ethnicity and age (less than 6 years, or more).
Intervention
After the first 4 weeks of prednisolone which was open label and common to both groups, participants were randomized to either:
Standard Course Treatment: a further 4 weeks of prednisolone at 40 mg/m2 on alternate days (total 8 weeks of prednisolone, total dose 2240 mg/m2), followed by 8 weeks of a placebo pill to complete 16 weeks.
Extended Course Treatment: a further 12 weeks of prednisolone starting at 60 mg/m2 every other day, and tapering by 10 mg/m2 every 2 weeks (total 16 weeks of prednisolone, total dose 3150 mg/m2).
Inclusion criteria
Age 1-14 years
First episode of idiopathic nephrotic syndrome (first morning urine protein to creatinine ratio or albumin to creatinine ratio > 200 mg/mmol, serum/plasma albumin level <25 g/L)
No previous treatment with steroids or other immunosuppressant
Exclusion criteria
History of poor adherence to treatment
Diseases other than minimal change glomerulonephritis where biopsy was performed
Known allergy to prednisolone
Participants were assessed at regular trial visits for a minimum of 24 months, and up to 48 months. As per routine clinical practice, parents would monitor albuminuria with Albustix, maintain a diary - and would contact the study site in case of a relapse, and their primary care doctor in case of other concerns.
Analysis and Outcomes
The primary outcome of the study was the time from starting open label prednisolone treatment to first relapse.
Secondary analyses included:
Prespecified subgroup analyses for age (5 years or less vs 6 years or more) and ethnicity (South Asian, white or other).
Relapse rate
Incidence of frequently relapsing nephrotic syndrome
Incidence of steroid dependent nephrotic syndrome
Use of second-line immunosuppressive drugs
Rates of adverse events
In addition, it included an economic analysis, with consideration of cost and quality of life driving the cost-effectiveness analysis.
The primary analysis was using the log rank test on Kaplan-Meier survival curves. A Cox proportional hazard regression analysis was done adjusting for ethnicity and age.
Given that in the standard group, there could be early relapses (in the period that the extended period was still on steroids) or steroid dependency (during same period), relapse events before 18 weeks were counted as having occurred at 18 weeks in them. Various sensitivity analyses were also conducted.
Funding
This study was funded by the National Institute for Health Research (NIHR) Health Technology Assessment and the pilot trial for this study was funded by Kidney Research UK and Kid’s Kidney Research.
Notably, there was patient participation in the design of the trial, with involvement of the UK Nephrotic Syndrome trust, and UK Renal Patient Support Group.
Results
In total, 237 children were randomized to either a standard course of treatment or an extended course of treatment between July 2011 and October 2014. This met the predetermined sample size for 80% power to detect a 20% absolute difference in relapse rate at 1 year, accounting for patient drop out. The authors noted that the trial participants accounted for around one third of all UK children who presented with steroid sensitive nephrotic syndrome over the trial period. From a pediatric nephrology trial perspective, this was a massive undertaking.
Figure 1 from Webb et al, BMJ 2019
Both the standard and extended course groups had similar baseline characteristics (Table 1). The mean age was 4.9 years and 20% of participants were of South Asian ethnicity. This is important as the authors pointed out that idiopathic nephrotic syndrome is 4-6 times more common in children of South Asian origin. Roughly two-thirds of participants were boys.
Table 1 from Webb et al, BMJ 2019
Roughly 80% of the participants had a relapse during the study.
The difference in time to first relapse between an extended (16 week) course of prednisolone treatment vs a standard (8 week) course of treatment was not significant (median 139 vs 87 days, log rank p = 0.28, hazard ratio 0.87, 95% confidence interval 0.65-1.17).
Figure 2 from Webb et al, BMJ 2019
No clear difference by age or ethnicity was found in subgroup analyses, although there may be weak evidence (p = 0.08 for interaction) that time to first relapse was increased in the extended course group for children aged 5 years or less compared to children aged 6 years or more (HR 0.72, 95% confidence interval 0.50-1.05). Interestingly, the intention to treat analysis excluded 14 participants (5 in extended group and 9 in standard group) who were randomized but were later withdrawn from the study after loss of corticosteroid sensitivity. The authors felt that this would not introduce bias as most of these dropouts occurred before the start of randomized treatment.
For secondary outcomes, there were no differences in the mean number of relapses or the proportion developing frequently relapsing or steroid dependent nephrotic syndrome or requiring 2nd line immunosuppressive treatment.
Table 2 from Webb et al, BMJ 2019
Adverse Effects
There were also no significant differences in adverse events except for patient-reported poor behavior, which was more common in the standard course group. However, when behavioral data was analyzed using the Achenbach child behavior checklist, there was no significant difference in behavior scores.
Table 3 from Webb et al, BMJ 2019
Cost-Effectiveness Analysis
Interestingly, the cost effectiveness analysis (presented in the supplementary appendices) found that the extended course of treatment was cheaper, by £3,260, than the standard course and produced a small gain in quality of life compared to the standard course, hence was dominant. This means that the extended steroid strategy was both more effective, and also was cheaper.
Supp Figure 4 from Webb et al, BMJ 2019: This shows that extended course of steroids, as compared to standard course, comes at a lower cost (lower half) and provides greater effectiveness (right half) and hence ‘dominates’
How did this happen? With the small difference in time to first relapse, there was a small difference in the QALYs, with the extended steroid strategy providing a higher QALY than standard care. Additionally, the early relapses triggered a few hospital and family doctor/practice nurse visits, a few hospital admissions, and higher use of expensive steroid alternatives, such as cyclosporin, cyclophosphamide, and mycophenolate.
Supp Table 3 from Webb et al, BMJ 2019: Shows the costs associated with the management in both groups. As can be seen, there is greater cost from doctor and hospital visits, as well as non-steroid medication use in the standard course group
Supp Table 4 from Webb et al, BMJ 2019: The small difference in early relapses perhaps drives a tiny effect in QALYs seen in this table.
Discussion
In summary, this study showed that an extended 16-week course of prednisolone for children with steroid sensitive nephrotic syndrome did not reduce the rate of relapse compared to a standard 8-week course. Interestingly, despite the longer course of prednisolone in the extended course group, no difference in the rate of adverse events was found between the two groups.
The major strength of this study is the broad inclusion criteria for children with steroid sensitive nephrotic syndrome, with the trial population encompassing about one-third of all children presenting with steroid sensitive nephrotic syndrome in the UK during the trial period according to the authors. There was also good representation of children from South Asian origin which formed 20% of recruited participants. The study design was also rigorous with double-blinding and a placebo control which helped to minimize bias.
The weaknesses of this study included the potential exclusion of younger children who were not able to take the trial drug which was provided as a crushable tablet. Despite attempting to enroll and randomize only children who were thought to be steroid sensitive, 14 of the included participants were later found to be steroid resistant after randomization. These children were withdrawn from the trial and not included in analyses.
Given the difference, one can always wonder if a larger trial would have resulted in a p value crossing the 0.05 dichotomanic threshold. A Bayesian interpretation of the same results would hence be somewhat different than the classical one presented.
The main outcome of the study suggests that extended course steroids delays the onset of relapse a little bit, without really changing the course of the disease in the long run. However, the small, non significant differences did translate into greater cost, with the extended course dominating the standard course in the economic analysis.
Conclusion
What happens next?
Does this change the needle in terms of how we practice? This could be interpreted either way - that an extended course of steroids, while not being superior, did not cause more adverse effects, while potentially reducing resource use.
Those preferring shorter (standard course) might say this is not an inferior choice - and the overall relapse rate is similar at about 80%.
Do we need any more trials?