- Email: [email protected]
Risk of relapse in steroid-sensitive nephrotic syndrome: Effect of stage of post-prednisone adrenocortical suppression
Risk of relapse in steroid-sensitive nephrotic syndrome: Effect of stage of post-prednisone adrenocortical suppression Glucocorticoid therapy has been effective in inducing remission in the majority o f children with primary nephrotic syndrome9 However, glucocorticoid-induced adrenocortical suppression is associated with early relapse, whereas normal adrenoeortieal function seems to delay the subsequent relapse, We analyzed 201 episodes o f prednisone treatment in 47 children with steroid-sensitive nephrotic syndrome by the life-table method. The effect o f partial cortisol substitution was studied as well. Adrenocortical function was determined by a two-hour A C T H test..4 response o f 50% to 100% o f the minimum normal response was diagnosed as "'moderate" suppression, and a response o f less than 50 % as "'severe. "" Post-prednisone adrenocortical function was normal in 99 episodes (49%). Adrenocortical suppression occurred in 102 episodes; o f these, 68 were considered moderate and 34 severe. Adrenocortical suppression increased both the initial relapse rate and the final risk o f a relapse. Severe suppression was always associated with a relapse, the longest remission time being 0.5 year. In moderate suppression, several long, relapse-free intervals were observed, but the risk o f relapse was still higher than in episodes with normal adrenocortieal function. Cortisol substitution possibly decreased the risk o f a relapse after severe adrenocortical suppression, but not after moderate suppression. (J PEDIATR 103:553, 1983)
S. Leisti, M.D., and O. Koskimies, M.D. HelsinkL Finland
MOST CHILDREN with primary nephrotic syndrome respond favorably to glucocorticoid treatment ~,2; we have classified such patients as having steroid-sensitive nephrotic syndrome. A well-known side effect of therapy with glucocorticoids is the temporary suppression of adrenocortical function? We previously studied the association between prednisone-induced hypocortisolism and the subsequent course of steroid-sensitive nephrotic syndrome in a number of tests of the hypothalamic-pituitary-adrenocortical axis function. 46 Children with post-prednisone adrenocortical suppression had a shorter first remission than children with normal cortisol secretion?. 6 Our preliminary results indicated that some of the early relapses in steroidsensitive nephrotic syndrome could be prevented by partially substituting cortisol for a short time in patients with
From Children's Hospital, University o f Helsinki. Supported by a grant from the Sigrid Jusblius Foundation. Reprint requests." Olli Koskimies, M.D.. Children's Hospital, University o f Helsinki, SF00290 Helsinki 29, Finland.
post-prednisone hypocortisolism.7 We report our findings from a study of 201 episodes of prednisone treatment in children with SSNS who were given a post-therapy ACTH test. Our purpose was to estimate the risk of a relapse, to see if this correlated with the stage of adrenocortical suppression, and to confirm the possible benefit of a partial, short-term cortisol substitution.
SUBJECTS AND METHODS We analyzed 201 episodes of prednisone treatment for either an initial episode or for a relapse of steroid-sensitive nephrotic syndrome in 47 children. This was the total number of episodes in which a post-prednisone ACTH test was given at Children's Hospital from January 1, 1974, to August 31, 1982. The total number of prednisone treatments given during this period was 267. The children, 40 boys and seven girls, were aged between 0.6 and 14.6 years at the onset of nephrotic syndrome. Sixteen patients received only the initial treatment, and remained healthy during the subsequent followup; the remaining children had from one to 22 relapses,
TheJournalofPEDlATRICS
553
554
Leisti and Koskirnies
The Journal of Pediatrics October 1983
o~ 100 J
~00 u..I
80
r--
,,.,J
=
1
80
m
60 i
II hi
L~L~l L~2~20
15
15
10
2
~
_6t/_. . . .
~ . . . . . . . . .
J
LL
60
,_,1 . . . .
O
!
05
.
. . . . .
--~-0
. . . . .
'.
15
I// . . 4 . . . 6. .
20
RELAPSE-FREE
1'0
8
TLNE,
/.,0
YEARS
Fig, I. Life-table analysis of 167 posttreatment follow-ups of episodes of prednisone treatment in children with steroid-sensitive nephrotic syndrome by normal and subnormal plasma cortisol responscs in post-prednisone two-hour ACTH test. There were 99 episodes with normal responses (continuous line) and 68 with subnormal responses (dashed line) at start of follow-up. Two time scales, separated by gap, are shown. Figures above lines give number of patients not in relapse at times indicated. Difference between curves is significant (P < 0.01).
Table. Distribution of 20l prednisone treatment episodes in 47 children with steroid-sensitive nephrotic syndrome according to extent of post-prednisone adrenocortical suppression and cortisol substitution ACTH test response
No cortisol substitution
Cortisol substitution
Normal Moderate suppression Severe suppression
99 52 16
0 16 18
with a mean of seven per patient. The mean duration of follow-up was 4.7 years (range 0,01 to 9.2 years). The criteria of primary nephrotic syndrome and of relapse were those established by the International Study of Kidney Disease in Children (ISKDC). s Steroid sensitivity was determined by the disappearance of proteinuria within 28 days of prednisone treatment. Most of the episodes were carried out according to the I S K D C schedule. 8 Daily prednisone therapy (60 m g / m z in three doses for four weeks during the first episode; until remission in all subsequent episodes) was followed by a four-week period of intermittent prednisone administration (four days without, three days with; 40 m g / m 2 in three doses)? The second period was later modified in some patients to alternate-day prednisone treatment (one morning dose) consisting of either a constant (40 n ~ m 2) or tapering dose (40 to t0 m g / m 2 in weekly 10 m g / m 2 decrements). The mode of therapy was not used 'as a variable in the subsequent analysis because of the small number of episodes with modification.
15
20
15
I
0 0
t,; 0.5 RELAPSE-
I 1.0 FREE
6
6
i 1.5
i 2.0
TIME.
YEARS
Fig. 2. Life-table analysis of 167 post-prednisone follow-ups, during first two years, of episodes of nephrotic syndrome by stages of post-prednisone adrenocortical function as determined by two-hour ACTH test. There were 99 episodes with normal responses (upper continuous line), 52 with moderately subnormal responses (middle dashed line), and 16 with severely subnormal responses (lower dashed line). Figures above lines give number of patients not in relapse at time indicated. Significant differences were found between curves for severe and normal responses (P < 0.01) and between those for severe and moderate responses (P < 0.05).
A two-hour A C T H test 4 was performed one to 12 days after the last dose of prednisone. A random third of the episodes with subnormal response were followed by a partial substitution of cortisol into the treatment regimen. 7 Ten m g / m 2 cortisol, divided into a morning and an afternoon dose, was given for 45 days, and then 5 m g / m 2 in a single morning dose for an additional 45 days. The criterion for adr~nocortical suppression was the detection of a subnormal two-hour plasma corfisol concentration ( - 9 5 % tolerance limit) in the A C T H test? The lower limit of the normal two-hour cortisol value in the A C T H test was 811 n m o l / L (29.4 ~ g / d l ) ? W e further divided the subnormal responses into moderate (two-hour cortisol 406 to 810 n m o l / L ) and severe (0 to 405 n m o l / L ) (Table). ,Relapse-free intervals in the different groups were analyzed by the life-table method (Figs. 1, 2, and 3). l~ Each interval until the next relapse was treated as an independent event, starting at the end of the treatment in one episode and stopping on the first day of treatment in the next, or on August 31, 1982, if the patient was still in remission. Tests of significance were calculated according to Peto et al) ~
Volume 103 Number 4
Relapse in nephrotic syndrome
100
,J
555
100~
i
A
B
~0'
80 t[1
60
60
i
I 40
.
20
~
-~
L'-I ,
....
20
2_____q_-~
....
2 O
t 0.5
I__ 1.0
115
~---
210
L_ 2
2
L
~ ........
0.5
1.0
1
RELAPSE
- FREE
~_~L]~-----
00
~.5
2!0
T IM E ,
YEARS
Fig. 3. Life-table analysis of effect of partial cortisol substitution on risk of relapse in steroid-sensitive nephrotic syndrome in children with post-prednisone adrenocortical suppression. Continuous lines indicate episodes with no cortisol substitution, and interrupted lines episodes with partial cortisol substitution for initial 0.25 year of follow-up. A, Episodes with moderate adrenocortical suppression. Unsubstituted, 52; substituted, 16. B, Episodes with severe adrenocortical suppression. Unsubstituted, 16; substituted, 18. Figures above lines give number of patients not in relapse at times indicated. Difference between curves in B not significant (P < 0.10).
RESULTS In 167 episodes not allocated for cortisol substitution that showed severely suppressed adrenocortical function (n = 16), no interval after treatment to relapse exceeded 0.5 years in any child (Fig. 4). In episodes that showed moderately suppressed adrenocortical function (n = 52), nine had a remission longer than 0.5 years, but in six of these nine remissions, adrenocortical response was almost in the normal range (Fig. 4). In episodes with a normal post-prednisone A C T H test response (n = 99), a relapsefree interval longer than 0.5 year was found 33 times and longer than one year 20 times. The incidence of relapses at 0.5 year, calculated from the actual number at risk for each group, was 66% for episodes with normal responses, 82% for episodes with moderately subnormal responses, and 100% for episodes with severely subnormal responses (P < 0.05). A t 1.0 year, the incidence of relapses was 79% for normal and 84% for moderately subnormal responses (this difference was not significant). When all the episodes with subnormal responses were analyzed together, there was a significant difference between the incidence of relapses at one year for this group compared with for the normal group, 88% and 79%, respectively (P < 0.05). The relapse-free interval curves by the life-table method for the normal (n = 99) and subnormal (n = 68) groups of post-prednisone A C T H test responses were significantly different (P < 0.01; Fig. 1). The relapse rate was very high in each group immediately after the prednisone treatment. The most noticeable difference between the groups was the much higher relapse rate in the subnormal response group during the initial 0.4 year. Rapid relapsing in the normal response group continued for 0.9 years. No relapse
J 2000 o o u o gO
o
1500 o
o 0 (D n.-
o
1000
o
g I
~
o
500
~; 8 -~ . . . .
u? klJ
:~
F--.
0
0.01
o
o
o'
o
oo
o
AA
o o
9
o cD cDo o o lo 9 CD O0 0 Oooo o o ~ ~ ~o 9
o ~
~
c~
o
9
9 9
AA
o
~ o o BoO o o~ o- - o o -o . . . . . . . . . . . o o o o oooo oo _
0.03005
_o
"- - - ' -
_
0.1
RELAPSE-
0.3
9
05 FREE
1
3 TIME.
5
10
YEARS
Fig. 4. Correlation of plasma cortisol response (two-hour plasma eortisol concentration) in post-prednisone two-hour ACTH tests with subsequent relapse-free time in 167 episodes of prednisone treatment in children with steroid-sensitive nephrotic syndrome. Time scale is logarithmic. Upper dashed line, lower limit of normal response (two-hour plasma eortisol = 811 nmol/L); lower dashed line (two-hour plasma cortisol = 405 nmol/L) divides episodes with subnormal response into those showing moderate (n = 52) and severe (n = 16) adrenocortical suppression, o Relapse; 9 remission.
occurred in either group after 1.5 years. The maximum absolute difference in relapse-free percentage, 30.5%, was seen at 0.15 to 0.19 years of follow-up. The gap then slowly narrowed to the final difference of 9.0% at 1.5 years. In the normal group, the relapse-free percentage after 1.5 years was 19.5%, compared with 10.5% in the subnormal group, indicating an almost twofold difference. Fig. 2 depicts a two-year analysis by the life-table method of the relapse-free interval according to the sever-
$$6
Leisti and Koskirnies
ity of the post-prednisone adrenocortical suppression. In episodes with moderately subnormal responses (n = 52), fast relapsing occurred during the initial 0.4 year. The two-year curve for this group was not significantly different from that for normal responses, even though a significant difference (P < 0.05) was present at 0.5 year. The final relapse-free percentage was 14%, or 5.5% tess than that for normal responses. In episodes with severely subnormal responses (n = 16), there was continuous rapid relapsing during the initial 0.5 year, and no permanent remissions. This curve differed significantly from that for moderately subnormal responses (P < 0.05) and from that for normal responses (P < 0.01). The effect of partial cortisol substitution on relapse-free interval in the groups with different post-prednisone ACTH test responses was also studied. When all the treatments with subnormal responses were analyzed together (curves not shown), no significant difference was detected between the episodes with (n = 34) or without (n = 68) substitution. The same was true between the two groups with moderately subnormal responses (16 substituted vs 52 not substituted) (Fig. 3, A), and with severely subnormal responses (P < 0,10; Fig. 3, B). DISCUSSION Adrenocortical suppression increased the risk of relapse in children with nephrotic syndrome, Both pretreatment and posttreatment suppression were predictive of early relapse, but posttreatment suppression had a stronger correlation with subsequent clinical course. 6 In the diagnosis and follow-up of prednisone-induced adrenocortical suppression, we prefer the ACTH test over the insulin test because of the absence of side effects, We chose the two-hour cortisol concentration as the sole criterion of response in the two-hour ACTH test, as it is the most accurate index of response with the least intraindividual variability.4 The intraindividual variability of the basal cortisol concentration is less precise than the two-hour cortisol response, and therefore may not be appropriate for the evaluation and follow-up of adrenocortical capacity, even though it does have some predictive value for the subsequent course of steroid-sensitive nephrotic syndrome.6, t~. ~2 We adopted the life-table method to analyze the risk of relapse by considering each individual episode of prednisone treatment of steroid-sensitive nephrotic syndrome as an independent event. We confirmed our preliminary finding of the significantly increased risk of relapse in children with post-prednisone adrenocortical suppression. s'6 This risk was also evident from the correlation plot presented in Fig. 4, but the life-table analysis was more
The Journal of Pediatrics October 1983
powerful in pointing out the difference in the kinetics of the relapse-rate between the episodes with normal and with suppressed adrenocortical capacity. Patients with subnormal adrenocortical function had an extremely rapid relapse rate during the initial 0.4 year; relapsing in this group then ceased. The rate of relapse, although fast, was considerably slower in children with normal function; in this group, however, relapsing continued for twice as long as in the group with subnormal function. It may be hypothesized that during the first 0.4 year, adrenocortical failure adds critically to a presently unidentified risk factor that is responsible for rapid relapsing during the first year. That adrenocortical suppression may also work as an independent risk factor is suggested by the fact that children with normal adrenocortical function have significantly more long remissions than children with adrenocortical suppression. We classified the suppression as moderate or severe by halving the subnormal range of the two-hour ACTH test response. A clear correlation between the extent of adrenocortical suppression and the relapsing was found by the life-table method. No patient with severely suppressed function stayed in remission for longer than 0.5 year. Patients with moderately suppressed adrenocortical function had a moderate risk of relapse when compared with patients with severe suppression and those with normal adrenocortical function. The life-table analysis did not show any significant difference between patients with moderate suppression and those with normal adrenocorticat function, but most patients with moderate suppression and long remissions had near normal A C T H test responses (Fig. 4). These responses were actually in the zone of diagnostic uncertainty when calculated on the basis of intraindividual variability of response? We previously suggested the use of partial cortisol substitution to prevent the risk of relapse after treatment in patients with adrenocortical suppression2 ~6 In a preliminary double-blin6Jcr0ssover study in 13 children with steroid-sensitive nephrotic syndrome, a 0.25 year partial cortisol substitution seemed to prevent a relapse at 0.5 year. 7 In our study, substitution had no effect after treatment in patients with moderately suppressed adrenocortical function, but in the 18 patients with severe suppression, two (and possibly three) stayed in long remission. Beoause. of the possible favorable effect after severe hypocortisolism, and also to prevent the risk of Addisonian crises, we give partial cortisol substitution to children with severe adrenocortical suppression only, until the A C T H test responses are normalized. Others have come to similar conclusions regarding cortisol substitution? 3 We thank Mrs. Riitta L6nnqvist for assistance.
Volume 103 Number 4 REFERENCES
1. Barnett HL, Schoeneman M, Bernstein J, Edelmann CM,Jr: Minimal change nephrotic syndrome. In Edelmann CM, editoi': Pediatric kidney disease. Boston, 1978, Little, Brown, p 695. 2. K0skimies O, Vilska J, Rapola J, Hallman N: Long-term outcome of primary nephrotic syndrome. Arch Dig Child 57:544, 1982. 3. Bethge H: Die steroidinduzierte Nebennierenunterfunktion: Pathogenese, Klinik, Diagnostik, Prophylaxe und Therapie. Klin Wochenschr 48:317, 1970. 4. Leisti S, Perheentupa J: Two-hour adrenocorticotropic hormone test: Accuracy in the evaluation of hypothalamicpituitary-adrenocortical axis. Pediatr Res 12:272, 1978. 5. Leisti S, Hallman N, Koskimies O, Perheentupa J, Rapola J, Vilska J: Association of postmedication hypocortisolism with early first relapse of idiopathic nephrotic syndrome. Lancet 2:795, 1977. 6. Leisti S, Vilska J, Hallman N: Adrenocortical insufficiency and relapsing in the idiopathic nephrotic syndrome in childhood. Pediatrics 60:334, 1977. 7. Leisti S, Koskimies O, Perheentupa J, Vilska J, Hallman N: Idiopathic nephroticsyndrome: Prevention of early relapse. Br Med J 1:892,1978.
Relapse in nephrotic syndrome
557
8. Abramowicz M, Barnett HL, Edelmann CM Jr, Greifer 1, Kobyashi O, Arneii GC, Barron BA, Gordillo H, Hallman N, Tiddens HA: Controlled trial of azathioprine in children with nephrotic syndrome. Lancet 1:959, 1970. 9. Leisti S, Ahonen P, Perheentupa J: The diagnosis and staging of hypocortisolism in progressing autoimmune adrenalitis. Pediatr Res (In press.) 10. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. 1I. Analysis and examples. Br J Cancer 35:1, 1977. 11. Moel D!, Kwun YA, Teitcher J: Hypothalamic-pituitaryadrenal (HPA) function in adrenocortical steroid treated minimal change nephrotie syndrome. Clin Nephrol 14:36, '1980. 12. L6pez-Gdmez JM, Canals M J, B/trcenas MC, Gdmez-Campderfi FJ, Rengel M, Morales MD, Lop6z-Ramos A, and Luque A: Prognostic value of the hypothalamic-pituitaryadrenal axis in the minimal change nephropathy. Kidney lnt 22:213, 1982 (abstr). 13. Arant BS, Singer SA, Bernstein J: Steroid-dependent nephrotic syndrome. J PEDJATR 100:328, 1982.
S. Leisti, M.D., and O. Koskimies, M.D. HelsinkL Finland
MOST CHILDREN with primary nephrotic syndrome respond favorably to glucocorticoid treatment ~,2; we have classified such patients as having steroid-sensitive nephrotic syndrome. A well-known side effect of therapy with glucocorticoids is the temporary suppression of adrenocortical function? We previously studied the association between prednisone-induced hypocortisolism and the subsequent course of steroid-sensitive nephrotic syndrome in a number of tests of the hypothalamic-pituitary-adrenocortical axis function. 46 Children with post-prednisone adrenocortical suppression had a shorter first remission than children with normal cortisol secretion?. 6 Our preliminary results indicated that some of the early relapses in steroidsensitive nephrotic syndrome could be prevented by partially substituting cortisol for a short time in patients with
From Children's Hospital, University o f Helsinki. Supported by a grant from the Sigrid Jusblius Foundation. Reprint requests." Olli Koskimies, M.D.. Children's Hospital, University o f Helsinki, SF00290 Helsinki 29, Finland.
post-prednisone hypocortisolism.7 We report our findings from a study of 201 episodes of prednisone treatment in children with SSNS who were given a post-therapy ACTH test. Our purpose was to estimate the risk of a relapse, to see if this correlated with the stage of adrenocortical suppression, and to confirm the possible benefit of a partial, short-term cortisol substitution.
SUBJECTS AND METHODS We analyzed 201 episodes of prednisone treatment for either an initial episode or for a relapse of steroid-sensitive nephrotic syndrome in 47 children. This was the total number of episodes in which a post-prednisone ACTH test was given at Children's Hospital from January 1, 1974, to August 31, 1982. The total number of prednisone treatments given during this period was 267. The children, 40 boys and seven girls, were aged between 0.6 and 14.6 years at the onset of nephrotic syndrome. Sixteen patients received only the initial treatment, and remained healthy during the subsequent followup; the remaining children had from one to 22 relapses,
TheJournalofPEDlATRICS
553
554
Leisti and Koskirnies
The Journal of Pediatrics October 1983
o~ 100 J
~00 u..I
80
r--
,,.,J
=
1
80
m
60 i
II hi
L~L~l L~2~20
15
15
10
2
~
_6t/_. . . .
~ . . . . . . . . .
J
LL
60
,_,1 . . . .
O
!
05
.
. . . . .
--~-0
. . . . .
'.
15
I// . . 4 . . . 6. .
20
RELAPSE-FREE
1'0
8
TLNE,
/.,0
YEARS
Fig, I. Life-table analysis of 167 posttreatment follow-ups of episodes of prednisone treatment in children with steroid-sensitive nephrotic syndrome by normal and subnormal plasma cortisol responscs in post-prednisone two-hour ACTH test. There were 99 episodes with normal responses (continuous line) and 68 with subnormal responses (dashed line) at start of follow-up. Two time scales, separated by gap, are shown. Figures above lines give number of patients not in relapse at times indicated. Difference between curves is significant (P < 0.01).
Table. Distribution of 20l prednisone treatment episodes in 47 children with steroid-sensitive nephrotic syndrome according to extent of post-prednisone adrenocortical suppression and cortisol substitution ACTH test response
No cortisol substitution
Cortisol substitution
Normal Moderate suppression Severe suppression
99 52 16
0 16 18
with a mean of seven per patient. The mean duration of follow-up was 4.7 years (range 0,01 to 9.2 years). The criteria of primary nephrotic syndrome and of relapse were those established by the International Study of Kidney Disease in Children (ISKDC). s Steroid sensitivity was determined by the disappearance of proteinuria within 28 days of prednisone treatment. Most of the episodes were carried out according to the I S K D C schedule. 8 Daily prednisone therapy (60 m g / m z in three doses for four weeks during the first episode; until remission in all subsequent episodes) was followed by a four-week period of intermittent prednisone administration (four days without, three days with; 40 m g / m 2 in three doses)? The second period was later modified in some patients to alternate-day prednisone treatment (one morning dose) consisting of either a constant (40 n ~ m 2) or tapering dose (40 to t0 m g / m 2 in weekly 10 m g / m 2 decrements). The mode of therapy was not used 'as a variable in the subsequent analysis because of the small number of episodes with modification.
15
20
15
I
0 0
t,; 0.5 RELAPSE-
I 1.0 FREE
6
6
i 1.5
i 2.0
TIME.
YEARS
Fig. 2. Life-table analysis of 167 post-prednisone follow-ups, during first two years, of episodes of nephrotic syndrome by stages of post-prednisone adrenocortical function as determined by two-hour ACTH test. There were 99 episodes with normal responses (upper continuous line), 52 with moderately subnormal responses (middle dashed line), and 16 with severely subnormal responses (lower dashed line). Figures above lines give number of patients not in relapse at time indicated. Significant differences were found between curves for severe and normal responses (P < 0.01) and between those for severe and moderate responses (P < 0.05).
A two-hour A C T H test 4 was performed one to 12 days after the last dose of prednisone. A random third of the episodes with subnormal response were followed by a partial substitution of cortisol into the treatment regimen. 7 Ten m g / m 2 cortisol, divided into a morning and an afternoon dose, was given for 45 days, and then 5 m g / m 2 in a single morning dose for an additional 45 days. The criterion for adr~nocortical suppression was the detection of a subnormal two-hour plasma corfisol concentration ( - 9 5 % tolerance limit) in the A C T H test? The lower limit of the normal two-hour cortisol value in the A C T H test was 811 n m o l / L (29.4 ~ g / d l ) ? W e further divided the subnormal responses into moderate (two-hour cortisol 406 to 810 n m o l / L ) and severe (0 to 405 n m o l / L ) (Table). ,Relapse-free intervals in the different groups were analyzed by the life-table method (Figs. 1, 2, and 3). l~ Each interval until the next relapse was treated as an independent event, starting at the end of the treatment in one episode and stopping on the first day of treatment in the next, or on August 31, 1982, if the patient was still in remission. Tests of significance were calculated according to Peto et al) ~
Volume 103 Number 4
Relapse in nephrotic syndrome
100
,J
555
100~
i
A
B
~0'
80 t[1
60
60
i
I 40
.
20
~
-~
L'-I ,
....
20
2_____q_-~
....
2 O
t 0.5
I__ 1.0
115
~---
210
L_ 2
2
L
~ ........
0.5
1.0
1
RELAPSE
- FREE
~_~L]~-----
00
~.5
2!0
T IM E ,
YEARS
Fig. 3. Life-table analysis of effect of partial cortisol substitution on risk of relapse in steroid-sensitive nephrotic syndrome in children with post-prednisone adrenocortical suppression. Continuous lines indicate episodes with no cortisol substitution, and interrupted lines episodes with partial cortisol substitution for initial 0.25 year of follow-up. A, Episodes with moderate adrenocortical suppression. Unsubstituted, 52; substituted, 16. B, Episodes with severe adrenocortical suppression. Unsubstituted, 16; substituted, 18. Figures above lines give number of patients not in relapse at times indicated. Difference between curves in B not significant (P < 0.10).
RESULTS In 167 episodes not allocated for cortisol substitution that showed severely suppressed adrenocortical function (n = 16), no interval after treatment to relapse exceeded 0.5 years in any child (Fig. 4). In episodes that showed moderately suppressed adrenocortical function (n = 52), nine had a remission longer than 0.5 years, but in six of these nine remissions, adrenocortical response was almost in the normal range (Fig. 4). In episodes with a normal post-prednisone A C T H test response (n = 99), a relapsefree interval longer than 0.5 year was found 33 times and longer than one year 20 times. The incidence of relapses at 0.5 year, calculated from the actual number at risk for each group, was 66% for episodes with normal responses, 82% for episodes with moderately subnormal responses, and 100% for episodes with severely subnormal responses (P < 0.05). A t 1.0 year, the incidence of relapses was 79% for normal and 84% for moderately subnormal responses (this difference was not significant). When all the episodes with subnormal responses were analyzed together, there was a significant difference between the incidence of relapses at one year for this group compared with for the normal group, 88% and 79%, respectively (P < 0.05). The relapse-free interval curves by the life-table method for the normal (n = 99) and subnormal (n = 68) groups of post-prednisone A C T H test responses were significantly different (P < 0.01; Fig. 1). The relapse rate was very high in each group immediately after the prednisone treatment. The most noticeable difference between the groups was the much higher relapse rate in the subnormal response group during the initial 0.4 year. Rapid relapsing in the normal response group continued for 0.9 years. No relapse
J 2000 o o u o gO
o
1500 o
o 0 (D n.-
o
1000
o
g I
~
o
500
~; 8 -~ . . . .
u? klJ
:~
F--.
0
0.01
o
o
o'
o
oo
o
AA
o o
9
o cD cDo o o lo 9 CD O0 0 Oooo o o ~ ~ ~o 9
o ~
~
c~
o
9
9 9
AA
o
~ o o BoO o o~ o- - o o -o . . . . . . . . . . . o o o o oooo oo _
0.03005
_o
"- - - ' -
_
0.1
RELAPSE-
0.3
9
05 FREE
1
3 TIME.
5
10
YEARS
Fig. 4. Correlation of plasma cortisol response (two-hour plasma eortisol concentration) in post-prednisone two-hour ACTH tests with subsequent relapse-free time in 167 episodes of prednisone treatment in children with steroid-sensitive nephrotic syndrome. Time scale is logarithmic. Upper dashed line, lower limit of normal response (two-hour plasma eortisol = 811 nmol/L); lower dashed line (two-hour plasma cortisol = 405 nmol/L) divides episodes with subnormal response into those showing moderate (n = 52) and severe (n = 16) adrenocortical suppression, o Relapse; 9 remission.
occurred in either group after 1.5 years. The maximum absolute difference in relapse-free percentage, 30.5%, was seen at 0.15 to 0.19 years of follow-up. The gap then slowly narrowed to the final difference of 9.0% at 1.5 years. In the normal group, the relapse-free percentage after 1.5 years was 19.5%, compared with 10.5% in the subnormal group, indicating an almost twofold difference. Fig. 2 depicts a two-year analysis by the life-table method of the relapse-free interval according to the sever-
$$6
Leisti and Koskirnies
ity of the post-prednisone adrenocortical suppression. In episodes with moderately subnormal responses (n = 52), fast relapsing occurred during the initial 0.4 year. The two-year curve for this group was not significantly different from that for normal responses, even though a significant difference (P < 0.05) was present at 0.5 year. The final relapse-free percentage was 14%, or 5.5% tess than that for normal responses. In episodes with severely subnormal responses (n = 16), there was continuous rapid relapsing during the initial 0.5 year, and no permanent remissions. This curve differed significantly from that for moderately subnormal responses (P < 0.05) and from that for normal responses (P < 0.01). The effect of partial cortisol substitution on relapse-free interval in the groups with different post-prednisone ACTH test responses was also studied. When all the treatments with subnormal responses were analyzed together (curves not shown), no significant difference was detected between the episodes with (n = 34) or without (n = 68) substitution. The same was true between the two groups with moderately subnormal responses (16 substituted vs 52 not substituted) (Fig. 3, A), and with severely subnormal responses (P < 0,10; Fig. 3, B). DISCUSSION Adrenocortical suppression increased the risk of relapse in children with nephrotic syndrome, Both pretreatment and posttreatment suppression were predictive of early relapse, but posttreatment suppression had a stronger correlation with subsequent clinical course. 6 In the diagnosis and follow-up of prednisone-induced adrenocortical suppression, we prefer the ACTH test over the insulin test because of the absence of side effects, We chose the two-hour cortisol concentration as the sole criterion of response in the two-hour ACTH test, as it is the most accurate index of response with the least intraindividual variability.4 The intraindividual variability of the basal cortisol concentration is less precise than the two-hour cortisol response, and therefore may not be appropriate for the evaluation and follow-up of adrenocortical capacity, even though it does have some predictive value for the subsequent course of steroid-sensitive nephrotic syndrome.6, t~. ~2 We adopted the life-table method to analyze the risk of relapse by considering each individual episode of prednisone treatment of steroid-sensitive nephrotic syndrome as an independent event. We confirmed our preliminary finding of the significantly increased risk of relapse in children with post-prednisone adrenocortical suppression. s'6 This risk was also evident from the correlation plot presented in Fig. 4, but the life-table analysis was more
The Journal of Pediatrics October 1983
powerful in pointing out the difference in the kinetics of the relapse-rate between the episodes with normal and with suppressed adrenocortical capacity. Patients with subnormal adrenocortical function had an extremely rapid relapse rate during the initial 0.4 year; relapsing in this group then ceased. The rate of relapse, although fast, was considerably slower in children with normal function; in this group, however, relapsing continued for twice as long as in the group with subnormal function. It may be hypothesized that during the first 0.4 year, adrenocortical failure adds critically to a presently unidentified risk factor that is responsible for rapid relapsing during the first year. That adrenocortical suppression may also work as an independent risk factor is suggested by the fact that children with normal adrenocortical function have significantly more long remissions than children with adrenocortical suppression. We classified the suppression as moderate or severe by halving the subnormal range of the two-hour ACTH test response. A clear correlation between the extent of adrenocortical suppression and the relapsing was found by the life-table method. No patient with severely suppressed function stayed in remission for longer than 0.5 year. Patients with moderately suppressed adrenocortical function had a moderate risk of relapse when compared with patients with severe suppression and those with normal adrenocortical function. The life-table analysis did not show any significant difference between patients with moderate suppression and those with normal adrenocorticat function, but most patients with moderate suppression and long remissions had near normal A C T H test responses (Fig. 4). These responses were actually in the zone of diagnostic uncertainty when calculated on the basis of intraindividual variability of response? We previously suggested the use of partial cortisol substitution to prevent the risk of relapse after treatment in patients with adrenocortical suppression2 ~6 In a preliminary double-blin6Jcr0ssover study in 13 children with steroid-sensitive nephrotic syndrome, a 0.25 year partial cortisol substitution seemed to prevent a relapse at 0.5 year. 7 In our study, substitution had no effect after treatment in patients with moderately suppressed adrenocortical function, but in the 18 patients with severe suppression, two (and possibly three) stayed in long remission. Beoause. of the possible favorable effect after severe hypocortisolism, and also to prevent the risk of Addisonian crises, we give partial cortisol substitution to children with severe adrenocortical suppression only, until the A C T H test responses are normalized. Others have come to similar conclusions regarding cortisol substitution? 3 We thank Mrs. Riitta L6nnqvist for assistance.
Volume 103 Number 4 REFERENCES
1. Barnett HL, Schoeneman M, Bernstein J, Edelmann CM,Jr: Minimal change nephrotic syndrome. In Edelmann CM, editoi': Pediatric kidney disease. Boston, 1978, Little, Brown, p 695. 2. K0skimies O, Vilska J, Rapola J, Hallman N: Long-term outcome of primary nephrotic syndrome. Arch Dig Child 57:544, 1982. 3. Bethge H: Die steroidinduzierte Nebennierenunterfunktion: Pathogenese, Klinik, Diagnostik, Prophylaxe und Therapie. Klin Wochenschr 48:317, 1970. 4. Leisti S, Perheentupa J: Two-hour adrenocorticotropic hormone test: Accuracy in the evaluation of hypothalamicpituitary-adrenocortical axis. Pediatr Res 12:272, 1978. 5. Leisti S, Hallman N, Koskimies O, Perheentupa J, Rapola J, Vilska J: Association of postmedication hypocortisolism with early first relapse of idiopathic nephrotic syndrome. Lancet 2:795, 1977. 6. Leisti S, Vilska J, Hallman N: Adrenocortical insufficiency and relapsing in the idiopathic nephrotic syndrome in childhood. Pediatrics 60:334, 1977. 7. Leisti S, Koskimies O, Perheentupa J, Vilska J, Hallman N: Idiopathic nephroticsyndrome: Prevention of early relapse. Br Med J 1:892,1978.
Relapse in nephrotic syndrome
557
8. Abramowicz M, Barnett HL, Edelmann CM Jr, Greifer 1, Kobyashi O, Arneii GC, Barron BA, Gordillo H, Hallman N, Tiddens HA: Controlled trial of azathioprine in children with nephrotic syndrome. Lancet 1:959, 1970. 9. Leisti S, Ahonen P, Perheentupa J: The diagnosis and staging of hypocortisolism in progressing autoimmune adrenalitis. Pediatr Res (In press.) 10. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. 1I. Analysis and examples. Br J Cancer 35:1, 1977. 11. Moel D!, Kwun YA, Teitcher J: Hypothalamic-pituitaryadrenal (HPA) function in adrenocortical steroid treated minimal change nephrotie syndrome. Clin Nephrol 14:36, '1980. 12. L6pez-Gdmez JM, Canals M J, B/trcenas MC, Gdmez-Campderfi FJ, Rengel M, Morales MD, Lop6z-Ramos A, and Luque A: Prognostic value of the hypothalamic-pituitaryadrenal axis in the minimal change nephropathy. Kidney lnt 22:213, 1982 (abstr). 13. Arant BS, Singer SA, Bernstein J: Steroid-dependent nephrotic syndrome. J PEDJATR 100:328, 1982.