http://www.kidney-international.
org
& 2014 International Society of Nephrology
see commentary on page 17
Extending initial prednisolone treatment in a
randomized control trial from 3 to 6 months
did not significantly influence the course of illness in
children with steroid-sensitive nephrotic syndrome
Aditi Sinha1, Abhijeet Saha2, Manish Kumar3, Sonia Sharma1, Kamran Afzal4, Amarjeet Mehta5,
Mani Kalaivani6, Pankaj Hari1 and Arvind Bagga1
1
Division of Nephrology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India; 2Department of Pediatrics,
Postgraduate Institute of Medical Education and Research, Ram Manohar Lohia Hospital, New Delhi, India; 3Department of Pediatrics,
Chacha Nehru Bal Chikitsalaya, New Delhi, India; 4Department of Pediatrics, Jawaharlal Nehru Medical College, Aligarh, India;
5
Department of Pediatrics, Sawai Man Singh Medical College, Jaipur, India and 6Department of Biostatistics, All India Institute of
Medical Sciences, New Delhi, India
While studies show that prolonged initial prednisone therapy
reduces the frequency of relapses in nephrotic syndrome,
they lack power and have risk of bias. In order to examine the
effect of prolonged therapy on frequency of relapses, we
conducted a blinded, 1:1 randomized, placebo-controlled
trial in 5 academic hospitals in India on 181 patients, 1–12
years old, with a first episode of steroid-sensitive nephrotic
syndrome. Following 12 weeks of standard therapy, in
random order, 92 patients received tapering prednisolone
while 89 received matching-placebo on alternate days for
the next 12 weeks. On intention-to-treat analyses, primary
outcome of number of relapses at 1 year was 1.26 in the
6-month group and 1.54 in the 3-month group (difference
0.28; 95% confidence interval (CI) 0.75, 0.19). Relative
relapse rate for 6- vs. 3-month therapy, adjusted for gender,
age, and time to initial remission, was 0.70 (95% CI 0.47–1.10).
Similar proportions of patients had sustained remission,
frequent relapses, and adverse effects due to steroids.
Adjusted hazard ratios for first relapse and frequent relapses
with prolonged therapy were 0.57 (95% CI, 0.36–1.07) and
1.01 (95% CI, 0.61–1.67), respectively. Thus, extending initial
prednisolone treatment from 3 to 6 months does not
influence the course of illness in children with nephrotic
syndrome. These findings have implications for guiding the
duration of therapy of nephrotic syndrome.
Kidney International (2015) 87, 217–224; doi:10.1038/ki.2014.240;
published online 16 July 2014
KEYWORDS: corticosteroid; idiopathic nephrotic syndrome; minimal change;
prednisone; randomized controlled trial
Correspondence: Arvind Bagga, Division of Nephrology, Department of
Pediatrics, All India Institute of Medical Sciences, New Delhi 110029, India.
E-mail: arvindbagga@hotmail.com
Received 4 January 2014; revised 25 May 2014; accepted 29 May 2014;
published online 16 July 2014
Kidney International (2015) 87, 217–224
clinical trial
Most children with idiopathic nephrotic syndrome show
remission of proteinuria following therapy with cortico-
steroids.1 Patients with multiple relapses are at risk for
serious complications and medication-associated adverse
effects. Although prospective studies suggest that extended
therapy for the initial episode of nephrotic syndrome leads to
sustained remission and reduced frequency of relapses, the
optimal dose and duration of treatment are not clear.2–6
Although therapy for 12 weeks is considered standard,7,8 data
from randomized trials suggest that there might be benefit
of extending therapy up to 6 months, without the risk
of adverse effects due to steroids.9–12 Data across studies
are relatively consistent, but all trials were open label and
had methodological concerns, with risk of bias and over-
estimation of the effect size. It is noteworthy that a recent
double-blind study from The Netherlands failed to show
benefits of prolonged duration of prednisolone intake from
12 to 24 weeks, using equal cumulative doses.13 In a previous
study on 45 patients randomly allocated to receive 8 and 16
weeks of initial therapy,14 we found that time to first relapse
was longer with the latter treatment, but it did not lead to
lower frequency of relapses.
This prospective blinded study examined the hypothesis
that prolongation of initial prednisolone therapy from
3 months to 6 months reduces the frequency of relapses in
patients with steroid-sensitive nephrotic syndrome. The present
study is the first placebo-controlled randomized clinical trial
reporting the efficacy of extending the dose and duration of
initial corticosteroid therapy in reducing the risk of future
relapses.
RESULTS
Patient description
Between July 2010 and May 2012, 255 patients were screened
for eligibility at five academic centers in north India; 74
217
clinical trial A Sinha et al.: Prednisone for nephrotic syndrome

patients were excluded at onset or during 3 months of
initial therapy with prednisolone (Figure 1). Reasons for
nonparticipation of patients screened at individual hospitals
are given in Supplementary Table 1 online; their age and sex
distribution was similar to those randomized (data not
shown). Following 3 months of standard therapy, 92 patients
were randomly allocated to receive additional prednisolone
for 12 weeks (6-month group) and 89 to placebo (3-month
group).
The clinical and laboratory characteristics of patients
randomized to the two groups were similar at onset and
at randomization (Supplementary Table 2 online, Table 1).
Initial corticosteroid therapy was delayed in 34 patients
owing to infections or hypovolemia (Supplementary Table 2
online). The time to remission was comparable, with 95%
patients in remission at a median of 20 days. Patients in the
two groups received similar dose and duration of standard
therapy (Table 1).
Intervention
All randomized patients except one in the 3-month group
began the allocated intervention. Trial medication was dis-
continued owing to relapse in 11 (12.0%) children allocated
to receive 12 weeks of therapy with prednisolone and
23 (26.1%) allocated to placebo (P ¼ 0.015) (Figure 1). This
resulted in intervention being administered for 81.4±
18.9 days in patients receiving prednisolone compared
with 75.2±21.6 days in those receiving placebo (P ¼ 0.045).
The cumulative initial dose of prednisolone was 3529.7±
398.7 mg/m2 in the 6-month group and 2791.7±286.6 mg/m2

255 Patients assessed for eligibility

36 Excluded
13 Staying far
12 Declined consent
8 Received prior steroid therapy
3 Impaired renal function

219 Enrolled, received standard therapy (3 months)

38 Excluded
20 Initial resistance
10 Relapsed
3 Received erratic therapy
5 Unable to contact

181 Randomized

92 Randomized to receive prednisolone 89 Randomized to receive placebo

(6-Month group) (3-Month group)
92 Received prednisolone as randomized 88 Received placebo as randomized
1 Did not return after randomization

3-Month follow-up 3-Month follow-up

91 Follow-up available 86 Follow-up available
0 Lost to follow-up 2 Lost to follow-up
1 Excluded due to steroid-resistant relapse 23 Discontinued intervention for treatment of
11 Discontinued intervention for treatment of steroid-sensitive relapse
steroid-sensitive (10) or -resistant (1) relapse

12-Month follow-up 12-Month follow-up

91 Follow-up for primary outcome available 83 Follow-up for primary outcome available
0 Lost to follow-up 5 Lost to follow-up
1 Excluded for steroid-resistant relapse

92 Included in intent-to-treat analysis 88 Included in intent-to-treat analysis

1 Missing data censored 5 Missing data censored

Figure 1 | Study flow.

218 Kidney International (2015) 87, 217–224

A Sinha et al.: Prednisone for nephrotic syndrome clinical trial

Table 1 | Patient characteristics at randomization

6-Month group, N ¼ 92 3-Month group, N ¼ 89
Age at onset, months 44.2 (34.2, 74.4) 42.4 (30.0–70.5)
Boys 56 (60.9%) 59 (67.1%)
Weight SDS  0.78 (  1.59, 0.04)  0.64 (  1.76, 0.08)
Height SDS  1.20 (  2.21,  0.44)  1.23 (  2.3,  0.32)
Body mass index SDS 0.72 (  0.28, 1.36) 0.37 (  0.12, 1.39)
Systolic blood pressure SDS 1.03 (0.37, 1.65) 0.88 (0.31, 1.64)
Diastolic blood pressure SDS 1.09 (0.61, 1.56) 1.17 (0.56, 1.67)
Hypertension 29 (31.5%) 25 (28.1%)
Estimated GFR, ml/min per 1.73 m2 92.4±34.2 (65.5, 107.9) 88.9±38.5 (62.4, 103.1)
Duration of standard therapy, days 84.3±4.3 (84, 85) 84.9±2.4 (84, 85)
Cumulative prednisolone, mg/m2 2784.9±261.9 (2614.9, 2921.5) 2791.7±286.6 (2601.9, 2952.4)
Days to remission 11.2±3.5 (9, 13) 11.5±4.4 (8, 13)
Abbreviations: GFR, glomerular filtration rate; SDS, standard deviation scores.
Values for continuous variables represent median (interquartile range) or mean±s.d. (interquartile range), as appropriate; two-sided P values calculated with w2 or t test (all
P40.1).

Table 2 | Outcomes at 12 months of follow-up

Relative risk Mean/percent risk
Outcome 6-Month group 3-Month group [95% CI] difference [95% CI] P
Intention-to-treat population N ¼ 92 N ¼ 88
Number of relapsesa 1.26±1.58 (0, 2) 1.54±1.59 (0, 3) —  0.28 [  0.75, 0.19] 0.24
Occurrence of relapse 49 (53.3%) 55 (62.5%) 0.85 [0.66, 1.09]  9.2 [  23.6, 5.2]% 0.21
Frequent relapses 35 (38.0%) 35 (39.8%) 0.96 [0.66, 1.38]  1.8 [  16.0, 12.5]% 0.81
Use of steroid-sparing agent 10 (10.9%) 15 (17.1%) 0.64 [0.30, 1.34]  6.2 [  16.3, 3.9]% 0.23
Cumulative prednisolone 2310.5±1921.1 (769.4, 3429.2) 1860.1±1928.2 (0, 3460.2) — 450.4 [  120.6, 1021.4] 0.12
received, mg/m2
Per-protocol population N ¼ 81 N ¼ 60
Number of relapsesb 0.93±1.15 (0, 2) 0.95±1.17 (0, 1.5) —  0.02 [  0.41, 0.37] 0.90
Occurrence of relapse 38 (46.9%) 31 (51.7%) 0.91 [0.65, 1.27]  4.8 [  21.5, 11.9]% 0.57
Frequent relapses 26 (32.1%) 15 (25.0%) 1.28 [0.75, 2.20] 7.1 [  7.8, 22.0]% 0.37
Use of alternative agent 2 (2.5%) 5 (8.3%) 0.30 [0.06, 1.48]  5.9 [  13.6, 1.9]% 0.11
Cumulative prednisolone 1938.8±1612.9 (754.7, 2896.5) 1015.9±1371.0 (0, 1213.8) — 922.9 [409.2, 1436.6] 0.005
received, mg/m2
Values represent n (percentage, %) or mean±s.d. (interquartile range); 95% confidence intervals (CI) are shown in square brackets.
a
The number of relapses was 1.27 and 1.55 per person-year in the 6-month and 3-month groups, respectively, based on 131 relapses during 1012 person-months and 116
relapses in 1094 person-months in the respective groups, with an incident rate ratio of 0.82 [95% CI 0.64, 1.05; P ¼ 0.13].
b
The number of relapses was 0.93 and 0.95 per person-year in the 6-month and 3-month groups, respectively, based on 75 relapses during 972 person-months and 57
relapses in 720 person-months in the respective groups, with an incident rate ratio of 0.97 [95% CI 0.69, 1.38; P ¼ 0.95].

in the 3-month group (mean difference 738.9 mg/m2; 95% CI
635.5–840.6 mg/m2; Po0.0001).
Participant flow
All 180 patients who received the allocated intervention were
included in the analyses of outcomes (Figure 1). The mean
duration of follow-up from randomization was 28.7±6.9
months in the 6-month group and 27.9±8.5 months in the
3-month groups. Follow-up for 12 months was available for
91 and 83 patients, respectively. Of 5 patients in the 3-month
group lost to follow-up between 1 and 6.5 months, 4 were
in sustained remission and one had a single relapse. One patient
in the 6-month group was diagnosed with late steroid resistance
2.5 months after randomization and was excluded. Outcomes
for these 6 patients were censored at last follow-up.
Primary outcome
Table 2 shows that the number of steroid-sensitive relapses
during the year following randomization were similar in
patients receiving 6 months vs. 3 months of therapy (mean
difference  0.28 relapses per year; 95% confidence interval,
CI  0.75, 0.19; P ¼ 0.24). The frequency of infection-
associated relapses was similar (0.76±1.30 vs. 0.96±1.06
relapses; mean difference  0.20, 95% CI  0.59, 0.12;
P ¼ 0.19). Poisson regression with the generalized estimating
equations approach, to account for clustering of relapses,
showed that the relative relapse rate was 0.86 (95% CI,
0.61–1.20; P ¼ 0.37) for patients treated for 6 months vs.
3 months, and 0.70 (95% CI 0.47–1.10; P ¼ 0.11) when
adjusted for sex, age, and time to remission (Supplementary
Table 3 online).
Secondary outcomes
Similar proportions of patients allocated to 6 months and
3 months of therapy relapsed during 1-year follow-up
(difference  9.2%; 95% CI  23.6%, 5.2%; P ¼ 0.21). The
proportions of patients with frequent relapses were compar-
able (difference  1.8%; 95% CI  16.0%, 12.5%; P ¼ 0.81).

Kidney International (2015) 87, 217–224 219

clinical trial A Sinha et al.: Prednisone for nephrotic syndrome

Assuming worst outcomes for patients lost to follow-up, Per protocol analysis
the proportions of patients with relapse (mean difference Forty (22.1%) patients either relapsed during intervention
 13.1%; 95% CI  27.9, 0.5%; P ¼ 0.067) and frequent (n ¼ 34) or were not followed up for 12 months (n ¼ 6).
relapses (mean difference  7.5%; 95% CI  21.8, 6.9%; Analysis after excluding these patients confirmed that relapse
P ¼ 0.31) continued to be comparable. Similar numbers of rates were similar in the two groups (Table 2). Secondary
patients required long-term prednisolone or steroid-sparing outcomes at 1-year follow-up were comparable (Table 2), except
agents at 12 months and at last follow-up (Supplementary higher cumulative prednisolone dose in patients allocated to 6
Table 4 online; Tables 2 and 3). The cumulative prednisolone months of therapy (mean difference 922.9 mg/m2; 95% CI
received during 12 months, as intervention and for therapy of 409.2, 1436.6; P ¼ 0.0005). Survival analyses showed similar
relapses, was comparable (mean difference 450.4 mg/m2; 95% proportions of patients in sustained remission or with frequent
CI  120.6, 1021.4; P ¼ 0.12). relapses (data not shown).
At least one relapse was observed during follow-up in 62
of 92 (67.4%) children in the 6-month group and in 61 of 88
Adverse events
(69.3%) children in the 3-month group. Figure 2 shows that the
Adverse effects related to steroid therapy were similar at
respective median time to first relapse was 9.2 months and 6.6
randomization, close of intervention, and at 12-month
months (log rank P ¼ 0.15). The hazard ratio (HR) for first
follow-up (Table 4). One patient with frequent relapses had
relapse with prolonged therapy, adjusted for sex, age, and time
to remission, was 0.57 (95% CI 0.36–1.07; P ¼ 0.11) (Supple-
mentary Table 5 online). Relapse rates were comparable at 1.00
2 years and at last follow-up (Table 3). Total relapses during --- Prednisolone (6-month group)
___ Placebo (3-month group)
follow-up were 2.34±2.39 in the 6-month group and 2.36± Survival free of relapse
0.75
2.19 in the 3-month group (mean difference  0.03; 95% CI Hazard ratio 0.77 [95% CI 0.44–1.55]
 0.65, 0.70; P ¼ 0.94). On Poisson regression, the relative Log rank P =0.15
relapse rate at last follow-up, adjusted for sex, age category, 0.50
and time to remission, was 0.82 (95% CI 0.57, 1.22; P ¼ 0.35).
At the last follow-up, 50.4% patients in the 6-month 0.25
group and 60.4% patients in the 3-month group had frequent
relapses. The time to frequent relapses was 23.0 months and
0.00
17.6 months in patients receiving prolonged therapy and 0 6 12 18 24 30
3 months of therapy, respectively (Figure 3; log rank Time since randomization, months
P ¼ 0.17). The HR for frequent relapses with prolonged Group
6-Month 92 61 (31) 43 (18) 35 (4) 25 (6) 15 (2)
therapy, adjusted for sex, age, and time to remission and to
3-Month 88 47 (39) 30 (15) 20 (6) 17 (1) 12 (0)
first relapse, was 1.01 (95% CI 0.61–1.67; P ¼ 0.96) (Supple-
mentary Table 5 online). The proportion of patients with Figure 2 | Relapse-free survival. The proportions with sustained
remission in patients treated for 6 months and 3 months were similar
steroid dependence was similar at 12 months (9.4 vs.11.2%), at 12 months (46.7 vs. 36.2%), at 24 months (34.1 vs. 26.8%), and at
24 months (18.6 vs.16.8%), and at the last follow-up last follow-up (28.4 vs. 26.8%). The panel shows the number of
(23.0 vs.16.8%; log rank P ¼ 0.71). patients at risk (number relapsed) at each time point.

Table 3 | Outcomes at 24 months and at last follow-up, by intention-to-treat analysis

Relative risk Mean/percent risk
Outcome 6-Month group 3-Month group [95% CI] difference [95% CI] P
24-Month follow-up N ¼ 72 N ¼ 67
Number of relapses per yeara 1.01±1.04 (0, 2) 1.13±1.01 (0, 2) —  0.12 [  0.47, 0.23] 0.48
Occurrence of relapse 48 (66.7%) 47 (70.2%) 0.95 [0.76, 1.19]  3.5% [  18.9, 12.0] 0.67
Frequent relapses 35 (48.6%) 36 (53.7%) 0.90 [0.65, 1.25]  5.1% [  21.7, 11.5] 0.55
Use of steroid-sparing agent 33 (45.8%) 33 (49.3%) 0.93 [0.66, 1.32]  3.4% [  20.0, 13.2] 0.69
Cumulative prednisolone received, mg/m2 3741.1±1932.6 (774.4, 5666.1) 3420.1±3142.0 (0, 5386.2) — 321.1 [  763.1, 1405.2] 0.56
Last follow-upb N ¼ 92 N ¼ 88
Number of relapses per yearc 0.98±1.01 (0, 1.7) 1.12±1.11 (0, 1.9) —  0.15 [  0.46, 0.16] 0.35
Occurrence of relapse 62 (67.4%) 61 (69.3%) 0.97 [0.80, 1.19]  1.9% [  15.5, 11.7] 0.78
Frequent relapses 44 (47.8%) 48 (54.6%) 0.88 [0.66, 1.17]  6.7% [  21.3, 7.9] 0.37
Alternative agent 16 (17.4%) 23 (26.1%) 0.67 [0.38, 1.17]  8.7% [  20.8, 3.3] 0.16
Values represent n (percentage, %) or mean±s.d. (interquartile range); 95% confidence intervals (CI) are shown in square brackets.
a
The number of relapses was 1.01 and 1.12 per person-year in the 6-month and 3-month groups, respectively, based on 141 relapses during 1680 person-months and 145
relapses in 1560 person-months in the respective groups, with an incident rate ratio of 0.90 [95% CI 0.72, 1.13; P ¼ 0.42].
b
Follow-up duration was 28.7±6.9 months in the 6-month and 27.9±8.5 months in the 3-month groups.
c
The number of relapses was 1.03 and 1.09 per person-year in the 6-month and 3-month groups, respectively, based on 215 relapses during 2507 person-months and 208
relapses in 2283 person-months in the respective groups, with an incident rate ratio of 0.94 [95% CI 0.78, 1.14; P ¼ 0.57].

220 Kidney International (2015) 87, 217–224

A Sinha et al.: Prednisone for nephrotic syndrome clinical trial

1.00 --- Prednisolone (6-month group) Table 4 | Corticosteroid-associated adverse effects and
Survival free of frequent relapses

___ Placebo (3-month group) serious infections during 12 months

0.75 6-Month group 3-Month group P
Hazards ratio 0.75 [95% CI 0.50–1.13]
Log rank P =0.17 Steroid-related
0.50 Hypertension
At close of intervention 18 (19.6%) 14 (15.9%) 0.52
At 12 months 18 (19.6%) 10 (11.4%) 0.13
0.25 Cushingoid appearance
At close of intervention 37 (40.2%) 37 (42.1%) 0.80
At 12 months 30 (32.6%) 34 (38.6%) 0.39
0.00 Hirsutism
0 6 12 18 24 30 At close of intervention 10 (11.0%) 8 (9.1%) 0.67
At 12 months 2 (2.2%) 1 (1.1%) 0.59
Time since randomization, months
Short stature
Group
At close of intervention 24 (26.1%) 23 (26.1%) 0.99
6-Month 92 81 (10) 56 (25) 47 (2) 35 (7) 22 (0)
At 12 months 15 (16.3%) 15 (17.1%) 0.89
3-Month 88 62 (24) 49 (10) 36 (8) 30 (3) 18 (3)
Obesity
Figure 3 | Survival free of frequent relapses. Proportions of At close of intervention 5 (5.4%) 5 (5.7%) 0.94
patients with frequent relapses in the 6-month and 3-month groups At 12 months 3 (3.3%) 5 (5.7) 0.43
were 38.4 and 40.4% at 12 months, 50.4 and 56.5% at 24 months, and Aggressive behavior
50.4 and 60.4% at last follow-up. The panel shows the number of At close of intervention 4 (4.3%) 4 (4.5%) 0.99
patients at risk (number with frequent relapses) at each time point. At 12 months 4 (4.3%) 4 (5.7%) 0.74

Episodes of severe infection

Pneumonia 12 (13.6%) 13 (14.1%) 0.92
posterior subcapsular cataract at 24 months. The rates Peritonitis 2 (2.2%) 0 0.16
Septic shock 1 (1.1%) 0 0.33
of serious infections requiring hospitalization (Table 4) and
Enteric fever 2 (2.2%) 0 0.16
minor infections were similar (Supplementary Table 6 online). Malaria 3 (3.3%) 1 (1.1%) 0.33
SDS for height and body mass index showed comparable scores Varicella 1 (1.1%) 1 (1.1%) 0.98
at follow-up (Supplementary Figure 1 online).

Post hoc analyses

Subgroup analyses showed that the duration of initial therapy
had no effect on the relapses during 12 months in boys
or girls, and in patients with early (p10 days) or delayed
remission (Supplementary Figure 2a online). Although
logistic regression showed decreasing probability of the first
relapse and of frequent relapses with increasing age at onset,
there was no threshold below which prolonged therapy was
beneficial (Supplementary Figure 3 online). Cox regression,
however, suggested that children younger than 3 years
benefitted from prolonged therapy (Supplementary Tables 5
and 7 online), with a lower risk of first relapse (Supplemen-
tary Figure 2b online) but not frequent relapses (Supplemen-
tary Figure 2c online). Poisson regression confirmed a higher
relative relapse rate in younger patients (Supplementary
Table 3 online).
As prednisolone was dosed by weight rather than by
surface area, the cumulative dose for standard (12-weeks)
therapy in patients p3 years old was lower than in older
children (2633.2±185.2 mg/m2 vs. 2880.1±276.6 mg/m2;
Po0.0001). However, the cumulative prednisolone dose did
not influence the time to first relapse (univariate HR 0.96;
95% CI 0.51, 1.79; P ¼ 0.89). The reduced risk of first relapse
in younger children with prolonged therapy was unchanged
when adjusted for prednisolone dose (HR 0.53; 95% CI 0.30,
0.93; P ¼ 0.027). In addition, age p3 years, but neither
prednisolone dose nor prolonged therapy, was associated
with an increased risk of first relapse (adjusted HR 1.61; 95%
CI 1.07, 2.40; P ¼ 0.021).
DISCUSSION
Results from this multicenter randomized placebo-controlled
trial show that prolongation of initial prednisolone therapy
from 3 months to 6 months does not reduce the frequency
of relapses in children with steroid-sensitive nephrotic
syndrome. Although the first relapse was postponed by
3 months, corresponding to the duration of extended
corticosteroid therapy, the likelihood of sustained remission
was similar at 1 year. The proportion of patients with
frequent relapses, an important measure of disease severity,
was comparable. There was no significant difference in
prednisolone requirement or the use of corticosteroid-
sparing agents during 12–24 months of follow-up. These
findings are contrary to results from a meta-analysis that
showed benefit of extending the initial duration of treatment
beyond 3 months.2
Information from large cohorts suggests that 50–80%
of patients with nephrotic syndrome have disease relapses,
of which one-half are frequent.15,16 Reduction of relapses
is therefore an important objective of therapy. On the basis
of a study by the Arbeitsgemeinschaft für Pädiatrische
Nephrologie that showed reduced relapse rates on prolong-
ing initial corticosteroid therapy from two to three months,4
multiple controlled trials have examined the utility of further
extending such therapy. Results of four studies,9–12 compa-
ring 3 months of therapy with 6 months of therapy in 382
patients, showed that the latter was associated with a lower
relapse rate (mean difference  0.44 relapses per year), as

Kidney International (2015) 87, 217–224 221

clinical trial
well as reduced risks for relapse (relative risk 0.57) and
frequent relapses (relative risk 0.55).2 A recent randomized
trial showed lower annual relapse rates (1.54 vs. 0.63 per year;
P ¼ 0.011) and high rates of sustained remission (29 vs. 76%)
in patients receiving 3 months vs. 5 months of therapy.17
Many of these studies had methodological concerns that
might have led to overestimation of the benefits of prolonged
therapy. None of the studies was blinded, allocation conceal-
ment was inadequate,10,11,17 and patients lost to follow-up
were excluded from analyses;17 two studies were never
published.11,12 Although the meta-analysis suggested that
both the duration18 and dose6,12 of initial corticosteroid
therapy determined the risk of relapses,2 a recent Dutch
multicenter randomized placebo-controlled trial failed to
show benefit of extending therapy from 12 to 24 weeks
without increasing the total dose.13
Although extension of initial corticosteroid therapy
would delay the first relapse, the intent is to alter the disease
course and to reduce the frequency of relapses. The primary
outcome, the number of relapses during 12 months of follow-
up, is an important marker of disease severity that determines
corticosteroid requirement. The relapse rates in the present
study were comparable to those reported in other rando-
mized controlled trials.4,13,17 Although the time to first
relapse was longer by 2.6 months in patients receiving
prolonged treatment, the benefit failed to translate into
reduced frequency of relapses at 1 year and subsequently.
Furthermore, Poisson regression using the generalized esti-
mating equations approach showed no significant difference
in relapse rates between groups. A significant (18.9%)
proportion of patients had to discontinue the intervention
(prolonged therapy or placebo for 12 weeks) owing to the
occurrence of relapses. To examine whether an early stoppage
of intervention may have attenuated the effect of prolonged
therapy, per protocol analyses was performed on patients
completing the entire 12-week intervention. The analyses
showed that relapse rates and the risk of frequent relapses
were similar in the two groups. Patients receiving 6 months
of initial therapy showed significantly higher steroid intakes
during the 1-year follow-up, negating the benefits expected
from the study.
An important outcome of studies examining the intensity
of initial therapy is its effect on preventing frequent
relapses.4,9,13,17 Previous studies that examined the efficacy
of prolonged versus standard initial therapy show that apart
from the immediate benefits of extended initial treatment
patients show significant reduction in the proportion of
frequent relapsers, reflecting in reduced requirement for
corticosteroids and/or alternative agents.4,5,10 These benefits
were not observed in the present study, and almost 40% of
patients in both groups had frequent relapses at 12 months of
follow-up.
Previous studies suggest that an early age of onset is a risk
factor for frequent relapses.16,19,20 Hiraoka et al.9 showed that
children younger than 4 years benefit from prolonged initial
therapy with higher rates of remission and lower rates of
222
A Sinha et al.: Prednisone for nephrotic syndrome
frequent relapses. Post-hoc analysis in the present study
showed that children aged 3 years or younger benefit from
prolonged therapy with reduced risk for first relapse, but not
for frequent relapses. As prednisolone dose was calculated by
weight (rather than surface area), we found that compared
with older children those p3 years of age received 8.6%
lower dose per m2 during the initial 12 weeks. Consistent
with data from national surveys,21 a significant proportion of
patients in the present study showed low weight and height
indices. However, it is likely that malnourished children
would receive less prednisolone whether dosed by weight or
by body surface area. Although administration of lower doses
is proposed as a risk factor for relapses,22 unadjusted and
adjusted analyses in our patients showed that young age
and not lower steroid dose was associated with an increased
HR for relapse. Findings on subgroup analysis underscore
the need for prospective studies to examine the efficacy of
prolonged therapy in young children.
Our study has multiple limitations. First, the trial was not
stratified for baseline variables, specifically age and sex that
may affect disease severity; subgroup analyses were therefore
post hoc. Randomization was not stratified by center, as their
contributions were difficult to predict. Although findings
on per protocol analyses confirmed those on intention-to-
treat, these were not specified a priori. Finally, sample size
calculation was based on 2.1 relapses/year; we observed 25%
fewer relapses in the standard (3-month) limb that reduced
the power to examine difference in relapse rates. Strengths of
the study are its blinded, placebo-controlled multicenter
design and a low risk of selection, performance, detection,
and selective reporting bias. Inclusion criteria were defined to
limit postrandomization exclusions, and outcomes were
assessed using widely accepted definitions. Baseline features
were well balanced, and there was a low rate of attrition.
Regular safety assessment through clinical and biochemical
monitoring was ensured. We believe that results from this
study are generalizable to patients with the first episode of
steroid-sensitive nephrotic syndrome.
Findings from this study suggest that the benefit of
extended initial therapy, beyond 3 months, appears to be
limited to the period while the steroids are being adminis-
tered. Among children with the first episode of nephrotic
syndrome, prolongation of initial prednisolone treatment to
6 months is not effective in modifying the course of the
disease, and reducing subsequent need for corticosteroids
and steroid-sparing agents.
MATERIALS AND METHODS
Trial design
This investigator-initiated, prospective, randomized, placebo-
controlled, parallel-group trial compared the efficacy of 6 months
vs. 3 months of initial prednisolone therapy in decreasing the
frequency of relapses in patients with steroid-sensitive nephrotic
syndrome. The trial was approved by Ethics Committees at each site,
registered at the Clinical Trials Registry of India (http://ctri.nic.in;
CTRI/2010/091/001095) and was conducted in accordance with the
Kidney International (2015) 87, 217–224
A Sinha et al.: Prednisone for nephrotic syndrome
original protocol. This study report complies with the Consolidated
Standards of Reporting Trials statement. The coordinating center
(All India Institute of Medical Sciences) conceived and designed the
study protocol. Investigators at the center vouch for the accuracy
and completeness in collation and analysis of data.
Participants
Patients, aged 1–12 years old, with a first episode of idiopathic
nephrotic syndrome, defined as nephrotic-range proteinuria (3–4 þ
proteinuria by dipstick, or spot urine protein to creatinine, Up/Uc
X2.0 mg/mg), hypoalbuminemia (albumin o2.5 g/dl), and edema
were eligible. Patients with an estimated glomerular filtration rate
o60 ml/min per 1.73 m2,23 known secondary cause, e.g., systemic
lupus, Henoch Schonlein purpura, or hepatitis B antigenemia, gross
hematuria, and residence more than 200 km away were excluded.
We did not include patients who had received prior therapy with
corticosteroids, ever for nephrotic syndrome or within the preceding
4 weeks for any indication. After informed written consent, eligible
patients received standard therapy with prednisolone at 2 mg/kg
daily for 6 weeks, followed by 1.5 mg/kg on alternate days for
6 weeks. Patients were followed up at 4, 8, 11, and 12 weeks of
therapy. Remission was defined as negative or trace proteinuria on
dipstick analysis for 3 consecutive days; relapse was the recurrence of
nephrotic-range proteinuria for 3 days. Patients who did not achieve
remission following 8 weeks of standard therapy (steroid resistance),
relapse during standard therapy, or medication noncompliance for
414 days were excluded.
Randomization and masking
At 11 weeks of standard therapy, eligible patients were reconsented
by co-investigators at each center and information was faxed to the
coordinating center for randomization. Procedures for randomiza-
tion, and packing and distribution of medications, were conducted
at this center by individuals who were not involved in trial
implementation. Allocation sequence was generated using Stata
version 10.1 (StataCorp 2008; Stata Statistical Software: Release 10;
College Station, TX: StataCorp LP), with patients randomly
assigned, in a 1:1 ratio in permuted blocks of four, to receive 3
months of additional prednisolone (6-months group) or 3 months
of placebo (3-month group). Prednisolone or placebo was given at a
dose of 1, 0.75, and 0.5 mg/kg on alternate days for 4 weeks each. An
external pharmacy manufactured the identical-appearing sugar-
coated tablets of prednisolone and placebo in strengths of 5 and
10 mg, and packaged them in matching blister packs of 10 tablets
each. Medication, sufficient to last 12 weeks for children weighing
up to 40 kg, was packed in similar containers and labeled with
unique serial numbers based on the randomization list, ensuring
allocation concealment.
Trial medication was couriered to the investigators during the
last week of standard therapy to ensure initiation of intervention
following 12 weeks of standard therapy. In the event of a relapse
during intervention, trial medication was discontinued and therapy
for relapse was initiated as described below. The investigators,
patients, and outcome assessors were blinded to the randomization
schedule. Masking was maintained during data analysis, following
which the randomization code was broken.
Follow-up
Parents were instructed to examine the first morning urine for
protein by dipstick and record details of relapses, medications, and
Kidney International (2015) 87, 217–224
clinical trial
infections. Follow-up visits were scheduled every month during
intervention, and then every 3 months. Records were reviewed at
each visit; physical examination included anthropometry, blood
pressure, and features of relapse, infection, or steroid adverse effects.
Hypertension was defined as systolic or diastolic blood pressure
X95th percentile for sex, age, and height.24 Blood counts and levels
of creatinine, albumin, cholesterol, glucose, and electrolytes were
measured every 6 months; eyes were examined annually.
Relapses were treated with prednisolone at 2 mg/kg/day until
remission, followed by 1.5 mg/kg on alternate days for 4 weeks. In
patients with frequent relapses, defined as two relapses in 6 months,
prednisolone was tapered to 0.5 mg/kg on alternate days and given
for 12–18 months. Those failing the above received the following
agent(s): (i) levamisole, 2–2.5 mg/kg on alternate days for 12–18
months; (ii) cyclophosphamide, 2 mg/kg/day for 12 weeks; (iii)
mycophenolate mofetil, 600–1000 mg/m2/day for 12–24 months; or
(iv) tacrolimus, 0.1–0.2 mg/kg/day for 12–24 months. We preferred
to use cyclophosphamide in patients with significant steroid toxi-
city, suggested by cataract, body mass index 42 SDS,25 or stage II
hypertension. Patients on long-term steroids received daily
supplements of calcium (250–500 mg) and vitamin D (200–400 U).
Hypertension was treated with enalapril or amlodipine.
Outcomes
The primary outcome was the numbers of steroid-sensitive relapses
during 12 months of follow-up. Secondary outcomes at 12 months
were (i) proportion of patients in sustained remission, (ii) propor-
tion with frequent relapses, (iii) cumulative prednisolone use
(mg/m2 per year), (iv) need for steroid-sparing therapies, and
(v) frequency and type of adverse events. We also recorded the time
from randomization to first relapse and time to frequent relapses,
the latter being the time to the second relapse in any 6-month
period. Safety assessments included clinical and laboratory evalua-
tion and monitoring for adverse events, with reports to the ethics
committees and review by the Data and Safety Monitoring Board.
Features of steroid toxicity were obesity (body mass index 42 SDS),
short stature (height o  2 SDS),25 hypertension, Cushingoid
appearance, parent-reported behavioral changes, and cataract or
glaucoma. Patients with late steroid resistance or with an estimated
glomerular filtration rate o60 ml/min per 1.73 m2 persisting for
more than 2 weeks were considered treatment failure and excluded.
Statistical analysis
Categorical data were analyzed with Pearson’s chi-square or Fisher’s
exact test. Continuous data were expressed as median (interquartile
range) or mean±s.d. and analyzed using rank-sum or unpaired
or paired Student’s t-tests. On the basis of a previous study,14 the
sample size required to detect 30% difference in the mean number
of relapses from 2.1±1.5 relapses/yr with 3 months of therapy to
1.5±1.5 relapses/yr with 6 months of therapy (mean difference of
0.6 relapses/year), at 80% power and a error of 5%, was 89 patients
per group (Stata version 11.0; StataCorp 2009). Data were analyzed
using Stata version 12.0 (StataCorp 2012). All tests were two-sided,
with Po0.05 considered significant.
The intention-to-treat population included patients who
returned for the first follow-up visit after randomization. Efficacy
analyses were performed with the last observation carried forward.
To handle missing values at 12 months, we performed worst
outcome analysis for categorical data and additionally presented
relapse rates as events per person-year. We also report per protocol
223
clinical trial
analyses on patients who (i) received the allocated intervention
for 12 weeks and (ii) were followed up for 12 months. To account
for clustering of relapses, Poisson regression was performed
using the generalized estimating equations approach with a log link
and autoregressive correlation matrix. Relative relapse rates
were estimated in relation to treatment and adjusted for age, sex,
and time to initial remission. We used the Kaplan–Meier method
with log rank test to compare time from randomization to first
relapse or frequent relapses. Cox proportional model was used to
obtain unadjusted and adjusted HR with 95% CI, and logistic
regression to calculate probability of relapse at different ages in
relation to the intervention. Subgroup analyses were exploratory
and not predefined; variables for subgroup or regression analyses
included age at onset (p3 and 43 years), time to remission
(p10 days and 410 days), and early first relapse (within 6 months
from randomization).26
DISCLOSURE
All the authors declared no competing interests.
ACKNOWLEDGMENTS
The study was funded by the Indian Council of Medical Research.
The sponsors had no role in the design and conduct of the study;
collection, management, analysis, and interpretation of the data;
preparation, review, or approval of the manuscript; and decision to
submit the manuscript for publication. Results of this study were
presented during the Landmark Clinical Trials session at the Congress
of the International Pediatric Nephrology Association in Shanghai
(2013) and annual meeting of the Indian Society of Pediatric
Nephrology in Bangalore (2013). Funding: Indian Council of Medical
Research. Registration: CTRI/2010/091/001095. Source of support:
Research Grant, Indian Council of Medical Research.
SUPPLEMENTARY MATERIAL
Table S1. Details of patients screened at individual centers.
Table S2. Patient characteristics at onset of nephrotic syndrome.
Table S3. Relative relapse rates at 12-month follow-up using Poisson
regression and the generalized estimating equations approach, with
an autoregressive correlation matrix.
Table S4. Details of therapy for frequent relapses.
Table S5. Hazards for first relapse and frequent relapses.
Table S6. Episodes of minor adverse events during 12 months of
follow-up.
Table S7. Hazards for relapses with prolonged therapy in the
younger subgroup.
Figure S1. Standard deviation scores (SDS) during follow-up for (a)
height and (b) body mass index.
Figure S2. Subgroup analysis.
Figure S3. Probability of disease relapse and frequent relapses at 12-
month follow-up according to age.
Supplementary material is linked to the online version of the paper at
http://www.nature.com/ki
REFERENCES
1. Tarshish P, Tobin JN, Bernstein J et al. Prognostic significance of the
early course of minimal change nephrotic syndrome: Report of the
International Study of Kidney Disease in Children. J Am Soc Nephrol 1997;
8: 769–776.
2. Hodson EM, Willis NS, Craig JC. Corticosteroid therapy for nephrotic
syndrome in children. Cochrane Database Syst Rev 2007; 4: CD001533.
224
A Sinha et al.: Prednisone for nephrotic syndrome
3. Jayantha UK. Comparison of ISKDC regime with a six month steroid regime
in the treatment of steroid sensitive nephrotic syndrome [abstract].
7th Asian Congress of Pediatric Nephrology, Singapore, 2000, Nov 1–4.
4. Ehrich JH, Brodehl J. Long versus standard prednisone therapy for initial
treatment of idiopathic nephrotic syndrome in children.
Arbeitsgemeinschaft fur Padiatrische Nephrologie. Eur J Pediatr 1993;
152: 357–361.
5. Ueda N, Chihara M, Kawaguchi S et al. Intermittent vs. long-term tapering
prednisolone for initial therapy in children with idiopathic nephrotic
syndrome. J Pediatr 1998; 112: 122–126.
6. Hiraoka M, Tsukahara H, Haruki S et al. Older boys benefit from higher
initial prednisolone therapy for nephrotic syndrome. The West Japan
Cooperative Study of Kidney Disease in Children. Kidney Int 2000; 58:
1247–1252.
7. Kidney Disease Improving Global Outcomes (KDIGO) Glomerulonephritis
Work Group.. KDIGO Clinical Practice Guideline for Glomerulonephritis.
Kidney Int 2012; Suppl 2: 139–274.
8. Bagga A, Ali U, Banerjee S et al. Indian Pediatric Nephrology Group.
Indian Academy of Pediatrics. Management of steroid sensitive nephrotic
syndrome: revised guidelines. Indian Pediatr 2008; 45: 203–214.
9. Hiraoka M, Tsukahara H, Matsubara K et al. West Japan Cooperative
Study Group of Kidney Disease in Children.. A randomized study of two
long course prednisolone regimens for nephrotic syndrome in children.
Am J Kidney Dis 2003; 41: 1155–1162.
10. Ksiazek J, Wyszynska T. Short versus long initial prednisone treatment in
steroid-sensitive nephrotic syndrome in children. Acta Paediatr 1995; 84:
889–893.
11. Gulati S, Ahmed M, Sharma RK et al. Comparison of abrupt withdrawal
versus slow tapering regimen of prednisolone therapy in the
management of first episode of steroid responsive childhood idio-
pathic nephrotic syndrome [abstract]. Nephrol Dialysis Transplant 2001;
16: A87.
12. Pecoraro C, Caropreso MR, Passaro G et al. Therapy of first episode of
steroid responsive nephrotic syndrome: a randomized controlled trial
[abstract]. Pediatr Nephrol 2004; 19: C72.
13. Teeninga N, Kist-van Holthe J, van Rijskwijk N et al. Extending
prednisolone therapy does not reduce relapse in childhood nephrotic
syndrome. J Am Soc Nephrol 2012; 24: 149–159.
14. Bagga A, Hari P, Srivastava RN. Prolonged versus standard prednisolone
therapy for initial episode of nephrotic syndrome. Pediatr Nephrol 1999;
13: 824–827.
15. Sinha A, Hari P, Sharma PK et al. Disease course in steroid sensitive
nephrotic syndrome. Indian Pediatr 2012; 49: 881–887.
16. Kabuki N, Okugawa T, Hayakawa H et al. Influence of age at onset on the
outcome of steroid-sensitive nephrotic syndrome. Pediatr Nephrol 1998;
12: 467–470.
17. Mishra OP, Thakur N, Mishra RN et al. Prolonged versus standard
prednisolone therapy for initial episode of idiopathic nephrotic
syndrome. J Nephrol 2012; 25: 394–400.
18. Satomura K, Yamaoka K, Shima M et al. Standard vs. low initial dose of
prednisolone therapy for first episodes of nephrotic syndrome in children
[abstract]. Pediatr Nephrol 2001; 16: C117.
19. Andersen RF, Thrane N, Noergaard K et al. Early age at debut is a
predictor of steroid-dependent and frequent relapsing nephrotic
syndrome. Pediatr Nephrol 2010; 25: 1299–1304.
20. Takeda A, Matsutani H, Niimura F et al. Risk factors for relapse in
childhood nephrotic syndrome. Pediatr Nephrol 1996; 10: 740–741.
21. NFHS-3, 2005–06: India: volume I. Mumbai: International Institute for
Population Sciences and Macro International, 2007.
22. Mehls O, Hoyer P. Dosing of glucocorticoids in nephrotic syndrome.
Pediatr Nephrol 2011; 26: 2095–2098.
23. Schwartz GJ, Muñoz A, Schneider MF et al. New equations to estimate
GFR in children with CKD. J Am Soc Nephrol 2009; 20: 629–637.
24. National High Blood Pressure Education Program Working Group on High
Blood Pressure in Children and Adolescents.. The fourth report on the
diagnosis, evaluation, and treatment of high blood pressure in children
and adolescents. Pediatrics 2004; 114: 555–576.
25. www.who.int/growthref/tools/en, accessed 31 December 2013.
26. Report of the International Study of Kidney Disease in Children. Early
identification of frequent relapsers among children with minimal change
nephrotic syndrome. A report of the International Study of Kidney
Disease in Children. J Pediatr 1982; 101: 514–518.
Kidney International (2015) 87, 217–224