- Email: [email protected]
Predicting the response to cytotoxic therapy for childhood nephrotic syndrome: Superiority of response to corticosteroid therapy over histopathologic patterns
Predicting the response to cytotoxic therapy for childhood nephrotic syndrome: Superiority of response to corticosteroid therapy over histopathologic patterns Seth L. Schulman, MD, Bruce A. Koiser, MD, Martin S. Polinsky, MD, R. Srinivasan, PhD, and H. Jorge Baluarte, MD From the Departments of Pediatrics and Mathematics, St. Christopher's Hospital for Children and Temple University School of Medicine, Philadelphia To determine the utility of steroid response in classifying childhood nephrotic syndrome, we reviewed 119 biopsies in 92 children a g e d 1 to 16 years who had been followed for a mean of 7.2 years. Steroid responses were classified as steroid resistant, steroid dependent, and frequent relapser as defined by the International Study of Kidney Disease in Children. Biopsy specimens were classified as showing focal glomerulosclerosiS (FSGS) in 39 children, as showing lipoid nephrosis in 28, and as questionable in another 25 with either focal global Sclerosis, IgM nephropathy, or mesangial prominence and tubular changes. A strong agreement (p <0.0!) was found between children whose FSGS was steroid resistant and children whose lipoid nephrosis resulted in frequent relapses. The length of the remission after therapy with chlorambucil or cyclophosphamide was determined in 84 children. A significantly shorter length of remission after cytotoxJc drug therapy (p <0.05) was identified for patients with FSGS versus those with lipoid nephrosis; this difference became more significant for steroid-resistant patients in comparison with those who were steroid dependent or were frequent relapsers (p <0.005). Among all steroid-resistant patients, those with FSGS had shorter remissions than patients with other histologic changes (p <0.001). The data suggest that patterns of response to corticosteroid therapy correlate with the histologic abnormality. Thus steroid-sensitive patients need not undergo renal biopsy before receiving cytotoxic drugs. Steroid-resistant patients would benefit from a biopsy, because the findings tend to predict the outcome. (J PEDIATR1988;113:996-1001)
In 1981 the International StudY of Kidney Disease in Children I evaluated 471 patients with nephrotic syndrome and determined that a renal biopsy was not necessary before initiating a course of corticosteroid therapy, because Presented in part at the 20th annual meeting of the American Society of Nephrology, Washington, D.C., December 1987. Submitted for publication May 9, 1988; accepted June 27, 1988. Reprint requests: Seth L. Schulman, MD, Section of Pediatric Nephrology, St. Christopher's Hospital for Children, 2600 N. Lawrence St., Philadelphia, PA 19133.
996
minimal change nephrotic syndrome could be identified in 92% of cases by the initial response to such therapy. However, the question of whether to perform a renal biopsy before initiation of cytotoxic drug therapy in the FSGS
Focal segmental glomerulosclerosis
I
Child with ster0id-sensitive but frequently relapsing nephrotic syndrome, or with steroid-dependent nephrotic syndrome, remains controversiaM We retrospectively evaluated 119 biopsies performed in
Volume 1t 3 Number 6
Table
Cytotoxic therapy for nephrotic syndrome
997
I. Population characteristics based on histologic study Lipoid nephrosis (n = 28)
Age (yr) Hypertension (>95th percentile for age) C..... <80 m!/min/l.73 m2 BSA Hematuria (>5 erythrocytes per high-power field)
4.4 -+ 2.9 0~ 0* 21"
Queslionable (n = 25)
3.5 -+ 2.6~ 2* 0* 4
FSGS (n = 39)
6.4 _+ 4.8 14 8 15
c ...... Creatinineclearance;BSA, body surfacearea. FSGS versus lipoidnephrosisor questionable: *p <0.05. ~fp<0.01. Sp <0.001.
92 children with nephrotic syndrome over a period of 15 years. An attempt was made to correlate patterns of responsiveness to corticosteroids with the histologic diagnosis. In addition, we evaluated the length of remission after therapy with cytotoxic agents, comparing the histologic diagnosis with the response to corticosteroids, to determine whether corticosteroid response is a better predictor of the outcome after cytotoxic therapy. METHODS
Between 1971 and 1986, a total of 119 biopsies were performed in 92 patients with nephrotic syndrome, aged 1 to 16 years. Nephrosis was defined as urinary protein excretion >--40 m g / h r / m 2 body surface area and hypoalbuminemia (serum albumin --<2.5 gm/dl). Steroid response patterns were defined as (1) frequently relapsing, when there were >--4 relapses per year (2) steroid dependent, when relapses occurred during alternate-day prednisone therapy or within 2 weeks after its discontinuation, and (3) steroid resistant, in patients in whom no response occurred after 8 weeks of corticosteroid therapy. All patients were initially treated with prednisone, 60 mg/m2/day for 28 days and then 40 m g / m 2 every other day for another 28 days, after which it was discontinued. Relapses were treated with 60 mg/m2/day of prednisone until 5 to 7 days after proteinuria was no longer detected. The patient then received 40 m g / m 2 every other day, after which the dose was tapered. The mean interval between the period when corticosteroids were started and a biopsy was performed was 2.1 +- 2.6 years (mean _+ SD) (range 2 months to i4 years), Patients had biopsies either when they were resistant to corticosteroid therapy or before a course of cytotoxic therapy if they demonstrated signs of toxicity to eorticosteroids. At times, repeat biopsies were performed when a second course of cytotoxic therapy was considered. All b!opsy specimens ~were i n t e r p r e t e d by the same pathologist. Focal segmental glomerulosclerosis was diagnosed by the presence of focal and segmental involvement of the glomeruli, Lipoid nephrosis was defined b y the
presence of glomeruli with a normal appearance on light microscopy or with only mild mesangial prominence, by immunofluorescence staining with minimal or no IgM deposits, and by the lack of electron microscopic findings other than epithelial cell effacement. Biopsy specimens were classified as "questionable," an indeterminate group, when focal global sclerosis or tubular atrophy with increased mesangiat prominence was present by light microscopy or when IgM deposits were identified by immunofluorescence in the absence of focal segmental lesions. Cytotoxic drug therapy was administered to 84 patients over a total of 56 days. Chlorambucil was given to 62 (74%) patients and cyclophosphamide to the other 22; the cumulative doses Of these agents were 12.6 _+ 5.8 mg/kg and 126 _+ 37 mg/kg, respectively. The choice of cytotoxic agent was determined by the attending nephrologist. The percent of patients who were relapse free for up to 60 months after cytotoxic drug therapy was evaluated by means of life table analysis3; for the purpose of this analysis, a relapse was defined as the redevelopment of nephrotic range proteinuria. Children were followed for a mean of 7.2 _+ 4.6 years (range 0.5 to 17.5) after the diagnosis of nephrotic syndrome was made. Statistical analysis was carried out by means of the Student t test and chi-square analysis. The degree of agreement between steroid patterns and biopsy results was estimated by means of Cohen's kappa statistic.4 The differences in remission rates between groups were analyzed with the use of Gehan's Wilcoxon test? RESULTS The mean age of all children was 5.0 + 3.9 years. There were 55 (60%) boys and 37 girls. On the initial biopsy, FSGS was diagnosed in 39 (42%) children and lipoid nephrosis in 28 (30%) children; the biopsy findings were considered questionable in the remaining 25 (27%) (Table I). Statistically significant differences for age at onset were
998
S c h u l m a n et al.
I00
The Journal o f Pediatrics December 1988
]
I00--
:l~.l
--
.~. . . . I
so
= FSGS
. . . . . =Q LN ........
\ i l
- - = S R
:.
..... FR .......... = S D -
80-
t..~
".......
..--,.
Z 0
(rJ
-
(0
Z
=E
~- -.~
LI.I
"-
60
0
% %,~
9. ~ ' ~
~
"%..~, . W
40
~-<'-:~''
~L"-"N
"-'-"-"-"-"-'-
N
40 ~
~0
. .............. ..................
\
LN vs FSGS p = . 0 5
20-
FSGS vs Q p = N S
SR v$ FR p<.OOb
LN vs 0 p = NS
SR vs SD p<.O05 SD vs FR p = N S
I I0
I 20
I 30 TIME
I 40
I 50
I 60
I
I
I
I
I
I
I0
20
30
40
50
60
TIME
(months)
(months)
Fig. 2 . Life table analysis in all children in our study, showing percentage of patients in remission by comparing patterns of response to corticosteroids. SR, Steroid resistant; FR, frequent relapses; SD, steroid dependent; NS, not significant.
Fig. t . Life table analysis in all children in our study, showing percentage of patients in remission by comparing histopathologic findings. LN, Lipoid nephrosis; Q, questionable, NS, not significant.
Table II. Morphologic transition of histologic changes Subsequent biopsy Original biopsy
n
Lipoid nephrosis
Questionable
FSGS
Lipoid nephrosis Questionable FSGS
5 4 8
1 0 0
1 0 0
3 4 8
found between all children with biopsies classified as FSGS or questionable. These differences occurred because girls with FSGS represented an older subgroup (8.1 + 5.1 years). This disparity in age was not evident for the boys. Abnormalities such as hypertension (>95th percentile for age and sex), 6 hematuria (>5 erythrocytes per highpower field), and decreased renal function, defined as a calculated creatinine clearance 7 <80 ml/min/1.73 m 2 body surface area, were present less frequently at the time of initial than at subsequent biopsies. When repeat biopsies
were also included, the majority of patients with hypertension and hematuria had FSGS, as did all patients with decreased renal function. The 17 repeat biopsy specimens demonstrated either the same findings or evidence of progression of the underlying disease (Table II). Six patients progressed to end-stage renal disease; initially, two had biopsy specimens considered questionable and four had FSGS. One patient whose biopsy specimen was categorized as questionable died of sepsis. A statistically significant agreement (p <0.01) was found between patients whose FSGS was steroid resistant
Volume 113 Number 6
Cytotoxic therapy for nephrotic syndrome
999
I0~
T a b l e III. Correlation of histologic pattern to steroid
response Histologic results = FSGS
Steroid pattern
Lipoid nephrosis
Questionable
FSGS
Frequent relapser Steroid dependent Steroid resistant
13 9 6
4 9 12
2 14 23
----~
and patients whose lipoid nephrosis resulted in frequent relapses, indicating that histopathologic changes correlate with response to corticosteroid therapy (Table III). Response to cytotoxic drugs was evaluated by comparing the duration of remissions achieved after treatment in patients with lipoid nephrosis, in those with FSGS, and in those with questionable histologic abnormalities (Fig9 1). Statistical significance was achieved only between patients with FSGS and those with tipoid nephrosis (p = 0.05). Conversely, the differences among these groups of patients became more significant when analyzed by patterns of eorticosteroid responsiveness '(Fig. 2). Significant differences were demonstrated between children who were steroid resistant versus either the steroid-dependent or the frequently relapsing groups (p <0.005). No statistical difference was demonstrated between steroid-dependent patients and frequent relapsers. Finally, when remission rates for steroid-resistant patients with the histologic diagnosis of FSGS were compared with those for patients with other histologic diagnoses, patients with FSGS had experienced a significantly shorter duration of remission (p <0.001) (Fig. 3). DISCUSSION The introduction of antibiotics and pharmacologic doses of corticosteroids has greatly reduced the mortality rate for nephrotic syndrome in children, which previously had been nearly 50%, 8 with most deaths attributable to infection. Unfortunately, steroid toxic effects, manifesting as obesity, growth retardation, cataracts, or aseptic necrosis, may limit the continued use of prednisone in the steroiddependent patient. The successful use of two cytotoxic agents, cyclophosphamide and chlorambucil, has been well documented, 9-11with one long-term study even demonstrating improvement in growth22 However, because of potential toxic effects, such as the development of bone marrow suppression, infection, sterility, and neoplasia, the use of these agents should be limited to the eorticosteroid-toxic patient. Keeping in mind that published data have not yet shown any improvement in the ultimate outcome of nephrotic syndrome after cytotoxic .drug therapy, the
z 0
= LN
-I- Q
p < .001
80
60
m
"'
40-
i
t. . . . . . . .
ZO
-
I
I
I
I
I
I0
ZO
30
40
50
TIME
(months)
Fi O. 3. Life table analysis in steroid-resistant children, showing
percentage of patients in remission by comparing histopathologic findings. LN, Lipoid nephrosis; Q, questionable.
clinician is faced with a difficult question: Should patients who cannot be maintained on a regimen of corticosteroids, either because of resistance or dependence with associated toxic effects, be allowed to remain nephrotic, or should they be treated with cytotoxic agents? In the past a renal biopsy has been considered to be indicated in patients with corticosteroid-responsive nephrotic syndrome before the use of cytotoxic agents, as well as in those who fail to respond to an initial course of prednisone/3.~4 Our data indicate that the prediction of response to cytotoxic therapy is likely to be more accurate when based on the initial corticosteroid response than when it is based on the patient's renal histopathologic findings; thus a renal biopsy may not be needed in steroid-responsive patients9 In support of this recommendation are the observations that the clinicopathologic correlations in idiopathic childhood nephrotic syndrome, as reported from other centers, are highly variable; the
1000
Schulman et al.
histologic pattern in patients with nephrotic syndrome may change with time; and, finally, there are inherent, although small, risks associated with performing a percutaneous renal biopsy. Available data differ regarding the accuracy of clinicopathologic correlations in childhood nephrosis. Studies by some authors 15"17 have documented a small likelihood of response to cytotoxic therapy and poor prognosis with FSGS; other reports 2' 18.19 are more favorable, with up to 51% of patients responding to immunosuppressive therapy. Similar ranges of variation have been reported for IgM nephropathy.16.20 Thus the evidence indicates that children with FSGS do not respond as well as those with either minimal change disease or those with mesangial hypercellularity or IgM deposition; these findings are similar to those of our study. However, they also demonstrate that a broad spectrum of responsiveness to therapy exists, making it difficult to justify the withholding of cytotoxic drug therapy in steroid-responsive children solely on the basis of histopathologic diagnosis. Tejani 2J and Goldzer et al. 22 described the overlap that may.be seen with regard to histologic findings in lipoid nephrosis, IgM nephropathy, and FSGS. Moreover, morphologic transitions between these three entities exist, and evolution to FSGS may occur, ~5,~7,23'24 supporting our findings. The situation is further complicated by the problem of obtaining representative samples in a disease with a predilection for beginning in juxtamedullary glomeruli. 25 Thus the reassurance based on early biopsy specimens may be premature, and in patients who experience a change in their clinical course a second tissue specimen may be required. It is difficult to determine the exact incidence of complications after renal biopsy in children. Major complications (e.g., death or the need for nephrectomy) are extremely uncommon. However, approximately 1% to 3% of patients undergoing percutaneous renal biopsy require a blood transfusion, and about 0.3% require surgical intervention. 13 Such morbidity will be avoided if biopsy can be deferred. Other studies predicting the likelihood of response to cytotoxic therapy have been reported. 9' 26,27 Garin et al. 26 studied the pattern of response in steroid-sensitive patients as a criterion for selecting children with nephrotic syndrome for cytotoxic therapy. They found that the response rates to cyclophosphamide differed significantly among frequently relapsing and steroid-dependent patients. Similar differences were found in the German collaborative prospective study, 9 which, in addition, found no advantage to the use of cyclophosphamide over that of chlorambucil. We did not find a significant difference between frequent relapsers and steroid-dependent children, possibly because
The Journal of Pediatrics December 1988
of differences in the doses of cytotoxic agents used. The response to higher doses of these agents improves significantly in steroid-dependent patients. 27 These studies also excluded histopathologic diagnoses other than minimal change nephrotic syndrome. Our study includes both steroid-resistant and steroid-dependent patients, and it indicates that steroid response patterns may be the best method of predicting the likelihood of response to cytotoxic drug therapy. Two conclusions may be drawn from our study. First, because of the aforementioned difficulties inherent in the pathologic interpretation of percutaneous biopsy material in FSGS, a focal progressive disease, we believe that the beneficial effect of cytotoxic therapy is better predicted by the pattern of response to prednisone than by the histologic classification. Such a conclusion justifies the use of cytotoxic agents before performing a renal biopsy in children with steroid-sensitive nephrotic syndrome and would spare a substantial group of children from the potential morbidity associated with this procedure. Second, children with steroid-resistant nephrosis benefit from a biopsy, because these children may have diseases other than FSGS not at all amenable to cytotoxic therapy, such as membranous nephropathy or membranoproliferative glomerulonephritis. ~ In addition, our data demonstrate that the combination of steroid resistance and FSGS is associated with a poor response to cytotoxic therapy. Because long-term follow-up is not available, cytotoxic therapy may be considered to prevent the morbidity associated with nephrosis, but a low remission rate should be expected in this subgroup. REFERENCES
1. International Study of Kidney Disease in Children. The primary nephrotic syndrome in children: identification of patients with minimal change nephrotic syndrome from initial response to prednisone. J PEDIATR1981;98:561-4. 2. Salcedo JR, Ruley EJ, Bock GH. Focal segmental glomerulosclerosis with idiopathic nephrotic syndrome. J PEDIATR 1983;102:325-6. 3. Cutler SJ, Ederer F. Maximum utilization of the life table method in analyzing survival. J Chronic Dis 1958;8:699712. 4. Fleis JL. Statistical methods for rates and proportions. New York: John Wiley & Sons, 1973. 5. Lee ET. Statistical methods for survival data dialysis. Belmont, Calif.: Lifetime Learning Publications, 1980. 6. Task Force on Blood Pressure Control in Children. Report of the second task force on blood pressure control in children. Pediatrics 1987;79:1-25. 7. Schwartz G J, Haycock GB, Edelmann CM Jr, Spitzer A. A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics 1976;8:259-63. 8. Arant BS, Singer SA, Bernstein J. Steroid-dependent nephrotic syndrome. J PEDIATR1982;100:328-33.
Volume 113 Number 6
9. Arbeitsgemeinschaft fiir P~idiatrische Nephrologie. Effect of eytotoxic drugs in frequently relapsing nephrotic syndrome with and without steroid dependence. N Engl J Med 1982; 306:451-4. 10, Balaarte H J, Hiner L, Gruskin AB. Chlorambucil dosage in frequently relapsing nephrotic syndrome: a controlled clinical trial. J PED~ATR 1978;92:295-8. 11. McDonald J, Murphy AV, Arniel GC. Long-term assessment of cyclophosphamide therapy for nephrosis in children. Lancet 1974;2:980-7. 12. Berns JS, Gaudio KM. Krassner LS, Anderson FP, Durante D, McDonald BM, Siegel NJ. Steroid-responsive nephrotic syndrome of childhood: a long-term study of clinical course, histopathology, efficacy of cyclophosphamide therapy and effects on growth. Am J Kidney Dis 1987;9:108-14. 13. Campos A, Vernier RL. In: Holliday MA, ed. Pediatric nephrology. Baltimore: Williams & Wilkins, 1987:330-7. 14. Gault MH, Muehrcke RC. Renal biopsy: current views and controversies. Nephron 1983;34:1-34. 15. Siegel N J, Gaudio KM, Krassner LS, McDonald BM, Anderson FP, Kashgarian M. Steroid-dependent nephrotic syndrome in children: histopathology and relapses after cyclophosphamide treatment. Kidney Int 1981 ;19:454-9. 16. Tejani A, Phadke K, Nicastri A, et al. Efficacy of cyclophosphamide in steroid-sensitive childhood nephrotic syndrome with different morphological lesions. Nephron 1985;41: 170-3. 17. Southwest Pediatric Nephrology Study Group. Focal segmental glomerulosclerosis in children with idiopathic nephrotic syndrome. Kidney Int 1985;27:442-9. 18. Mongeau J, Corneille L, Robitaille P, O'Regan S, Delletier
Cytotoxic therapy for nephrotic syndrome
19.
20.
21. 22. 23. 24.
25.
26.
27.
10 0 1
M. Primary nephrosis in childhood associated with focal glomerular sclerosis: is long-term prognosis that severe? Kidney Int 1981;20:743-6. Arbus GS, Poucell S, Bacheyie GS, Baumal R. Focal segmental glomerulosclerosis with idiopathic nephrotic syndrome: three types of clinical response. J PEDIATR 1982; 101:40-5. Vilches AR, Turner DR, Cameron JS, Ogg CS, Chantler C, Williams DG. Significance of mesangial IgM deposition in "minimal change" nephrotic syndrome. Lab Invest 1982; 46:10-5. Tejani A. Relapsing nephrotic syndrome. Nephron 1987; 45:81-5. Goldzer RC, Sweet J, Cotran RS. Focal segmental glomerulosclerosis. Ann Rev Med 1984;35:429-49. Tejani A. Morphological transition in minimal change nephrotic syndrome. Nephron 1985;39:157-9. Jennette JC. Evolution of mesangial IgM nephropathy into focal segmental glomerulosclerosis. Am J Nephrol 1981; 1:222. Habib R, Kleinkneckt C. The primary nephrotic syndrome of childhood: classification and clinicopathologic study of 406 cases. Pathol Annu 1971;6:417-74. Garin EH, Pryor ND, Fennell RS, Richard GA. Pattern of response to prednisone in idiopathic minimal lesion nephrotic syndrome as a criterion in selecting patients for eyclosphosphamide therapy. J PEDIATR,1978;92:304-8. Arbeitsgemeinschaft fiir P/idiatrisehe Nephrologie. Cyclophosphamide treatment of steroid-dependent nephrotic syndrome: comparison of eight week with 12 week course. Arch Dis Child 1987;62:1102-6.
In 1981 the International StudY of Kidney Disease in Children I evaluated 471 patients with nephrotic syndrome and determined that a renal biopsy was not necessary before initiating a course of corticosteroid therapy, because Presented in part at the 20th annual meeting of the American Society of Nephrology, Washington, D.C., December 1987. Submitted for publication May 9, 1988; accepted June 27, 1988. Reprint requests: Seth L. Schulman, MD, Section of Pediatric Nephrology, St. Christopher's Hospital for Children, 2600 N. Lawrence St., Philadelphia, PA 19133.
996
minimal change nephrotic syndrome could be identified in 92% of cases by the initial response to such therapy. However, the question of whether to perform a renal biopsy before initiation of cytotoxic drug therapy in the FSGS
Focal segmental glomerulosclerosis
I
Child with ster0id-sensitive but frequently relapsing nephrotic syndrome, or with steroid-dependent nephrotic syndrome, remains controversiaM We retrospectively evaluated 119 biopsies performed in
Volume 1t 3 Number 6
Table
Cytotoxic therapy for nephrotic syndrome
997
I. Population characteristics based on histologic study Lipoid nephrosis (n = 28)
Age (yr) Hypertension (>95th percentile for age) C..... <80 m!/min/l.73 m2 BSA Hematuria (>5 erythrocytes per high-power field)
4.4 -+ 2.9 0~ 0* 21"
Queslionable (n = 25)
3.5 -+ 2.6~ 2* 0* 4
FSGS (n = 39)
6.4 _+ 4.8 14 8 15
c ...... Creatinineclearance;BSA, body surfacearea. FSGS versus lipoidnephrosisor questionable: *p <0.05. ~fp<0.01. Sp <0.001.
92 children with nephrotic syndrome over a period of 15 years. An attempt was made to correlate patterns of responsiveness to corticosteroids with the histologic diagnosis. In addition, we evaluated the length of remission after therapy with cytotoxic agents, comparing the histologic diagnosis with the response to corticosteroids, to determine whether corticosteroid response is a better predictor of the outcome after cytotoxic therapy. METHODS
Between 1971 and 1986, a total of 119 biopsies were performed in 92 patients with nephrotic syndrome, aged 1 to 16 years. Nephrosis was defined as urinary protein excretion >--40 m g / h r / m 2 body surface area and hypoalbuminemia (serum albumin --<2.5 gm/dl). Steroid response patterns were defined as (1) frequently relapsing, when there were >--4 relapses per year (2) steroid dependent, when relapses occurred during alternate-day prednisone therapy or within 2 weeks after its discontinuation, and (3) steroid resistant, in patients in whom no response occurred after 8 weeks of corticosteroid therapy. All patients were initially treated with prednisone, 60 mg/m2/day for 28 days and then 40 m g / m 2 every other day for another 28 days, after which it was discontinued. Relapses were treated with 60 mg/m2/day of prednisone until 5 to 7 days after proteinuria was no longer detected. The patient then received 40 m g / m 2 every other day, after which the dose was tapered. The mean interval between the period when corticosteroids were started and a biopsy was performed was 2.1 +- 2.6 years (mean _+ SD) (range 2 months to i4 years), Patients had biopsies either when they were resistant to corticosteroid therapy or before a course of cytotoxic therapy if they demonstrated signs of toxicity to eorticosteroids. At times, repeat biopsies were performed when a second course of cytotoxic therapy was considered. All b!opsy specimens ~were i n t e r p r e t e d by the same pathologist. Focal segmental glomerulosclerosis was diagnosed by the presence of focal and segmental involvement of the glomeruli, Lipoid nephrosis was defined b y the
presence of glomeruli with a normal appearance on light microscopy or with only mild mesangial prominence, by immunofluorescence staining with minimal or no IgM deposits, and by the lack of electron microscopic findings other than epithelial cell effacement. Biopsy specimens were classified as "questionable," an indeterminate group, when focal global sclerosis or tubular atrophy with increased mesangiat prominence was present by light microscopy or when IgM deposits were identified by immunofluorescence in the absence of focal segmental lesions. Cytotoxic drug therapy was administered to 84 patients over a total of 56 days. Chlorambucil was given to 62 (74%) patients and cyclophosphamide to the other 22; the cumulative doses Of these agents were 12.6 _+ 5.8 mg/kg and 126 _+ 37 mg/kg, respectively. The choice of cytotoxic agent was determined by the attending nephrologist. The percent of patients who were relapse free for up to 60 months after cytotoxic drug therapy was evaluated by means of life table analysis3; for the purpose of this analysis, a relapse was defined as the redevelopment of nephrotic range proteinuria. Children were followed for a mean of 7.2 _+ 4.6 years (range 0.5 to 17.5) after the diagnosis of nephrotic syndrome was made. Statistical analysis was carried out by means of the Student t test and chi-square analysis. The degree of agreement between steroid patterns and biopsy results was estimated by means of Cohen's kappa statistic.4 The differences in remission rates between groups were analyzed with the use of Gehan's Wilcoxon test? RESULTS The mean age of all children was 5.0 + 3.9 years. There were 55 (60%) boys and 37 girls. On the initial biopsy, FSGS was diagnosed in 39 (42%) children and lipoid nephrosis in 28 (30%) children; the biopsy findings were considered questionable in the remaining 25 (27%) (Table I). Statistically significant differences for age at onset were
998
S c h u l m a n et al.
I00
The Journal o f Pediatrics December 1988
]
I00--
:l~.l
--
.~. . . . I
so
= FSGS
. . . . . =Q LN ........
\ i l
- - = S R
:.
..... FR .......... = S D -
80-
t..~
".......
..--,.
Z 0
(rJ
-
(0
Z
=E
~- -.~
LI.I
"-
60
0
% %,~
9. ~ ' ~
~
"%..~, . W
40
~-<'-:~''
~L"-"N
"-'-"-"-"-"-'-
N
40 ~
~0
. .............. ..................
\
LN vs FSGS p = . 0 5
20-
FSGS vs Q p = N S
SR v$ FR p<.OOb
LN vs 0 p = NS
SR vs SD p<.O05 SD vs FR p = N S
I I0
I 20
I 30 TIME
I 40
I 50
I 60
I
I
I
I
I
I
I0
20
30
40
50
60
TIME
(months)
(months)
Fig. 2 . Life table analysis in all children in our study, showing percentage of patients in remission by comparing patterns of response to corticosteroids. SR, Steroid resistant; FR, frequent relapses; SD, steroid dependent; NS, not significant.
Fig. t . Life table analysis in all children in our study, showing percentage of patients in remission by comparing histopathologic findings. LN, Lipoid nephrosis; Q, questionable, NS, not significant.
Table II. Morphologic transition of histologic changes Subsequent biopsy Original biopsy
n
Lipoid nephrosis
Questionable
FSGS
Lipoid nephrosis Questionable FSGS
5 4 8
1 0 0
1 0 0
3 4 8
found between all children with biopsies classified as FSGS or questionable. These differences occurred because girls with FSGS represented an older subgroup (8.1 + 5.1 years). This disparity in age was not evident for the boys. Abnormalities such as hypertension (>95th percentile for age and sex), 6 hematuria (>5 erythrocytes per highpower field), and decreased renal function, defined as a calculated creatinine clearance 7 <80 ml/min/1.73 m 2 body surface area, were present less frequently at the time of initial than at subsequent biopsies. When repeat biopsies
were also included, the majority of patients with hypertension and hematuria had FSGS, as did all patients with decreased renal function. The 17 repeat biopsy specimens demonstrated either the same findings or evidence of progression of the underlying disease (Table II). Six patients progressed to end-stage renal disease; initially, two had biopsy specimens considered questionable and four had FSGS. One patient whose biopsy specimen was categorized as questionable died of sepsis. A statistically significant agreement (p <0.01) was found between patients whose FSGS was steroid resistant
Volume 113 Number 6
Cytotoxic therapy for nephrotic syndrome
999
I0~
T a b l e III. Correlation of histologic pattern to steroid
response Histologic results = FSGS
Steroid pattern
Lipoid nephrosis
Questionable
FSGS
Frequent relapser Steroid dependent Steroid resistant
13 9 6
4 9 12
2 14 23
----~
and patients whose lipoid nephrosis resulted in frequent relapses, indicating that histopathologic changes correlate with response to corticosteroid therapy (Table III). Response to cytotoxic drugs was evaluated by comparing the duration of remissions achieved after treatment in patients with lipoid nephrosis, in those with FSGS, and in those with questionable histologic abnormalities (Fig9 1). Statistical significance was achieved only between patients with FSGS and those with tipoid nephrosis (p = 0.05). Conversely, the differences among these groups of patients became more significant when analyzed by patterns of eorticosteroid responsiveness '(Fig. 2). Significant differences were demonstrated between children who were steroid resistant versus either the steroid-dependent or the frequently relapsing groups (p <0.005). No statistical difference was demonstrated between steroid-dependent patients and frequent relapsers. Finally, when remission rates for steroid-resistant patients with the histologic diagnosis of FSGS were compared with those for patients with other histologic diagnoses, patients with FSGS had experienced a significantly shorter duration of remission (p <0.001) (Fig. 3). DISCUSSION The introduction of antibiotics and pharmacologic doses of corticosteroids has greatly reduced the mortality rate for nephrotic syndrome in children, which previously had been nearly 50%, 8 with most deaths attributable to infection. Unfortunately, steroid toxic effects, manifesting as obesity, growth retardation, cataracts, or aseptic necrosis, may limit the continued use of prednisone in the steroiddependent patient. The successful use of two cytotoxic agents, cyclophosphamide and chlorambucil, has been well documented, 9-11with one long-term study even demonstrating improvement in growth22 However, because of potential toxic effects, such as the development of bone marrow suppression, infection, sterility, and neoplasia, the use of these agents should be limited to the eorticosteroid-toxic patient. Keeping in mind that published data have not yet shown any improvement in the ultimate outcome of nephrotic syndrome after cytotoxic .drug therapy, the
z 0
= LN
-I- Q
p < .001
80
60
m
"'
40-
i
t. . . . . . . .
ZO
-
I
I
I
I
I
I0
ZO
30
40
50
TIME
(months)
Fi O. 3. Life table analysis in steroid-resistant children, showing
percentage of patients in remission by comparing histopathologic findings. LN, Lipoid nephrosis; Q, questionable.
clinician is faced with a difficult question: Should patients who cannot be maintained on a regimen of corticosteroids, either because of resistance or dependence with associated toxic effects, be allowed to remain nephrotic, or should they be treated with cytotoxic agents? In the past a renal biopsy has been considered to be indicated in patients with corticosteroid-responsive nephrotic syndrome before the use of cytotoxic agents, as well as in those who fail to respond to an initial course of prednisone/3.~4 Our data indicate that the prediction of response to cytotoxic therapy is likely to be more accurate when based on the initial corticosteroid response than when it is based on the patient's renal histopathologic findings; thus a renal biopsy may not be needed in steroid-responsive patients9 In support of this recommendation are the observations that the clinicopathologic correlations in idiopathic childhood nephrotic syndrome, as reported from other centers, are highly variable; the
1000
Schulman et al.
histologic pattern in patients with nephrotic syndrome may change with time; and, finally, there are inherent, although small, risks associated with performing a percutaneous renal biopsy. Available data differ regarding the accuracy of clinicopathologic correlations in childhood nephrosis. Studies by some authors 15"17 have documented a small likelihood of response to cytotoxic therapy and poor prognosis with FSGS; other reports 2' 18.19 are more favorable, with up to 51% of patients responding to immunosuppressive therapy. Similar ranges of variation have been reported for IgM nephropathy.16.20 Thus the evidence indicates that children with FSGS do not respond as well as those with either minimal change disease or those with mesangial hypercellularity or IgM deposition; these findings are similar to those of our study. However, they also demonstrate that a broad spectrum of responsiveness to therapy exists, making it difficult to justify the withholding of cytotoxic drug therapy in steroid-responsive children solely on the basis of histopathologic diagnosis. Tejani 2J and Goldzer et al. 22 described the overlap that may.be seen with regard to histologic findings in lipoid nephrosis, IgM nephropathy, and FSGS. Moreover, morphologic transitions between these three entities exist, and evolution to FSGS may occur, ~5,~7,23'24 supporting our findings. The situation is further complicated by the problem of obtaining representative samples in a disease with a predilection for beginning in juxtamedullary glomeruli. 25 Thus the reassurance based on early biopsy specimens may be premature, and in patients who experience a change in their clinical course a second tissue specimen may be required. It is difficult to determine the exact incidence of complications after renal biopsy in children. Major complications (e.g., death or the need for nephrectomy) are extremely uncommon. However, approximately 1% to 3% of patients undergoing percutaneous renal biopsy require a blood transfusion, and about 0.3% require surgical intervention. 13 Such morbidity will be avoided if biopsy can be deferred. Other studies predicting the likelihood of response to cytotoxic therapy have been reported. 9' 26,27 Garin et al. 26 studied the pattern of response in steroid-sensitive patients as a criterion for selecting children with nephrotic syndrome for cytotoxic therapy. They found that the response rates to cyclophosphamide differed significantly among frequently relapsing and steroid-dependent patients. Similar differences were found in the German collaborative prospective study, 9 which, in addition, found no advantage to the use of cyclophosphamide over that of chlorambucil. We did not find a significant difference between frequent relapsers and steroid-dependent children, possibly because
The Journal of Pediatrics December 1988
of differences in the doses of cytotoxic agents used. The response to higher doses of these agents improves significantly in steroid-dependent patients. 27 These studies also excluded histopathologic diagnoses other than minimal change nephrotic syndrome. Our study includes both steroid-resistant and steroid-dependent patients, and it indicates that steroid response patterns may be the best method of predicting the likelihood of response to cytotoxic drug therapy. Two conclusions may be drawn from our study. First, because of the aforementioned difficulties inherent in the pathologic interpretation of percutaneous biopsy material in FSGS, a focal progressive disease, we believe that the beneficial effect of cytotoxic therapy is better predicted by the pattern of response to prednisone than by the histologic classification. Such a conclusion justifies the use of cytotoxic agents before performing a renal biopsy in children with steroid-sensitive nephrotic syndrome and would spare a substantial group of children from the potential morbidity associated with this procedure. Second, children with steroid-resistant nephrosis benefit from a biopsy, because these children may have diseases other than FSGS not at all amenable to cytotoxic therapy, such as membranous nephropathy or membranoproliferative glomerulonephritis. ~ In addition, our data demonstrate that the combination of steroid resistance and FSGS is associated with a poor response to cytotoxic therapy. Because long-term follow-up is not available, cytotoxic therapy may be considered to prevent the morbidity associated with nephrosis, but a low remission rate should be expected in this subgroup. REFERENCES
1. International Study of Kidney Disease in Children. The primary nephrotic syndrome in children: identification of patients with minimal change nephrotic syndrome from initial response to prednisone. J PEDIATR1981;98:561-4. 2. Salcedo JR, Ruley EJ, Bock GH. Focal segmental glomerulosclerosis with idiopathic nephrotic syndrome. J PEDIATR 1983;102:325-6. 3. Cutler SJ, Ederer F. Maximum utilization of the life table method in analyzing survival. J Chronic Dis 1958;8:699712. 4. Fleis JL. Statistical methods for rates and proportions. New York: John Wiley & Sons, 1973. 5. Lee ET. Statistical methods for survival data dialysis. Belmont, Calif.: Lifetime Learning Publications, 1980. 6. Task Force on Blood Pressure Control in Children. Report of the second task force on blood pressure control in children. Pediatrics 1987;79:1-25. 7. Schwartz G J, Haycock GB, Edelmann CM Jr, Spitzer A. A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics 1976;8:259-63. 8. Arant BS, Singer SA, Bernstein J. Steroid-dependent nephrotic syndrome. J PEDIATR1982;100:328-33.
Volume 113 Number 6
9. Arbeitsgemeinschaft fiir P~idiatrische Nephrologie. Effect of eytotoxic drugs in frequently relapsing nephrotic syndrome with and without steroid dependence. N Engl J Med 1982; 306:451-4. 10, Balaarte H J, Hiner L, Gruskin AB. Chlorambucil dosage in frequently relapsing nephrotic syndrome: a controlled clinical trial. J PED~ATR 1978;92:295-8. 11. McDonald J, Murphy AV, Arniel GC. Long-term assessment of cyclophosphamide therapy for nephrosis in children. Lancet 1974;2:980-7. 12. Berns JS, Gaudio KM. Krassner LS, Anderson FP, Durante D, McDonald BM, Siegel NJ. Steroid-responsive nephrotic syndrome of childhood: a long-term study of clinical course, histopathology, efficacy of cyclophosphamide therapy and effects on growth. Am J Kidney Dis 1987;9:108-14. 13. Campos A, Vernier RL. In: Holliday MA, ed. Pediatric nephrology. Baltimore: Williams & Wilkins, 1987:330-7. 14. Gault MH, Muehrcke RC. Renal biopsy: current views and controversies. Nephron 1983;34:1-34. 15. Siegel N J, Gaudio KM, Krassner LS, McDonald BM, Anderson FP, Kashgarian M. Steroid-dependent nephrotic syndrome in children: histopathology and relapses after cyclophosphamide treatment. Kidney Int 1981 ;19:454-9. 16. Tejani A, Phadke K, Nicastri A, et al. Efficacy of cyclophosphamide in steroid-sensitive childhood nephrotic syndrome with different morphological lesions. Nephron 1985;41: 170-3. 17. Southwest Pediatric Nephrology Study Group. Focal segmental glomerulosclerosis in children with idiopathic nephrotic syndrome. Kidney Int 1985;27:442-9. 18. Mongeau J, Corneille L, Robitaille P, O'Regan S, Delletier
Cytotoxic therapy for nephrotic syndrome
19.
20.
21. 22. 23. 24.
25.
26.
27.
10 0 1
M. Primary nephrosis in childhood associated with focal glomerular sclerosis: is long-term prognosis that severe? Kidney Int 1981;20:743-6. Arbus GS, Poucell S, Bacheyie GS, Baumal R. Focal segmental glomerulosclerosis with idiopathic nephrotic syndrome: three types of clinical response. J PEDIATR 1982; 101:40-5. Vilches AR, Turner DR, Cameron JS, Ogg CS, Chantler C, Williams DG. Significance of mesangial IgM deposition in "minimal change" nephrotic syndrome. Lab Invest 1982; 46:10-5. Tejani A. Relapsing nephrotic syndrome. Nephron 1987; 45:81-5. Goldzer RC, Sweet J, Cotran RS. Focal segmental glomerulosclerosis. Ann Rev Med 1984;35:429-49. Tejani A. Morphological transition in minimal change nephrotic syndrome. Nephron 1985;39:157-9. Jennette JC. Evolution of mesangial IgM nephropathy into focal segmental glomerulosclerosis. Am J Nephrol 1981; 1:222. Habib R, Kleinkneckt C. The primary nephrotic syndrome of childhood: classification and clinicopathologic study of 406 cases. Pathol Annu 1971;6:417-74. Garin EH, Pryor ND, Fennell RS, Richard GA. Pattern of response to prednisone in idiopathic minimal lesion nephrotic syndrome as a criterion in selecting patients for eyclosphosphamide therapy. J PEDIATR,1978;92:304-8. Arbeitsgemeinschaft fiir P/idiatrisehe Nephrologie. Cyclophosphamide treatment of steroid-dependent nephrotic syndrome: comparison of eight week with 12 week course. Arch Dis Child 1987;62:1102-6.