- Email: [email protected]
Steroid-sensitive nephrotic syndrome: From childhood to adulthood
Steroid-Sensitive Nephrotic Syndrome: From Childhood to Adulthood Fadi Fakhouri, MD, Nathalie Bocquet, MD, Pierre Taupin, MD, Claire Presne, MD, Marie-France Gagnadoux, MD, Paul Landais, MD, PhD, Philippe Lesavre, MD, Dominique Chauveau, MD, Bertrand Knebelmann, MD, PhD, Michel Broyer, MD, Jean-Pierre Gru¨nfeld, MD, and Patrick Niaudet, MD ● Background: The clinical presentation, treatment, and outcome of steroid-sensitive nephrotic syndrome (SSNS) during childhood have been extensively studied. Conversely, few data regarding the outcome in adulthood of childhood SSNS have been published previously. We undertook to conduct a retrospective study of the outcome in adulthood of a large cohort of patients diagnosed with an SSNS during childhood. Methods: We identified all children born between 1970 and 1975 who had been admitted to our institution for an SSNS. Data regarding the outcome in adulthood of these patients were obtained through mailed questionnaires or phone calls to patients and/or their parents or through attending physicians. Results: One hundred seventeen patients were identified. Data regarding the outcome of SSNS in adulthood were available for 102 patients (87.2%). Forty-three patients (42.2%) experienced at least one relapse of nephrotic syndrome in adulthood. By univariate analysis, young age at onset (<6 years) and more severe disease in childhood, indicated by a greater number of relapses (12.9 for adulthood relapsers versus 5.4 for adulthood nonrelapsers; P < 0.0001) and more frequent use of immunosuppressors (74.4% versus 31.6%; P < 0.0001) or cyclosporine (42.9% versus 7.3%; P < 0.0001) were predictive of the occurrence of SSNS relapse in adulthood. Conversely, relapse rate in the first 6 months of disease was not predictive of further relapses in adulthood. By multivariate analysis, only number of relapses during childhood was predictive of adulthood relapses (P < 0.0058). Long-term side effects of steroids were found in 44.2% of adulthood relapsers; the most frequent were osteoporosis and excess weight. Conclusion: The incidence of childhood SSNS relapses in adulthood was relatively high in our study. Further studies are required to assess long-term complications in adults with relapses and a history of prolonged steroid and immunosuppressor use. Am J Kidney Dis 41: 550-557. © 2003 by the National Kidney Foundation, Inc. INDEX WORDS: Steroid-sensitive nephrotic syndrome (SSNS); childhood; relapse and adulthood.
S
TEROID-SENSITIVE nephrotic syndrome (SSNS) is the most frequent glomerular disease encountered during childhood.1 Pathophysiological characteristics of SSNS, most often related to minimal change disease, remain elusive. Previous studies have extensively assessed the clinical presentation, treatment, and outcome of this disease during childhood.2-5 SSNS relapses in up to 70% of children, with most of these patients frequent relapsers or steroid dependent.2 Immunosuppressive or immunomodulator drugs, including cyclophosphamide,4 chlorambucil,5 cyclosporine A,6 or levamisole,7 have proved effective as steroid-sparing agents From the Departments of Nephrology, Pediatric Nephrology, and Biostatistics, Assistance Publique-Hoˆpitaux de Paris, Hoˆpital Necker-Enfants Malades, Paris, France. Received June 22, 2002; accepted in revised form October 10, 2002. Address reprint requests to Fadi Fakhouri, MD, Service de Ne´phrologie, Hoˆpital Necker, 149 rue de Se`vres, 75743 Paris Cedex 15, France. E-mail: [email protected] © 2003 by the National Kidney Foundation, Inc. 0272-6386/03/4103-0003$30.00/0 doi:10.1053/ajkd.2003.50116 550
in frequently relapsing patients. Renal prognosis is usually favorable in these children despite their frequent relapses. Few data regarding the outcome in adulthood of childhood SSNS are available.8,9 Thus, the risk for relapse and incidence of treatment-related complications in adulthood remain largely unknown. This report is the first extensive study regarding the outcome in adulthood of a large group of patients diagnosed with SSNS during childhood. PATIENTS AND METHODS
Patients Using a computerized file, we identified all children: (1) with an SSNS, (2) born between 1970 and 1975 (thus aged ⬎ 25 years at the time of the study), and (3) who had been followed up (as inpatients or outpatients) in the Department of Paediatric Nephrology at the Hoˆpital NeckerEnfants Malades (Paris, France). Pediatric medical records of these patients were reviewed to collect relevant data. Information regarding the outcome of patients in adulthood (ⱖ18 years) was obtained from mailed questionnaires and phone calls to the patients and/or their parents, from adult nephrologists or general practitioners, and from medical records at Hoˆpital Necker-Enfants Malades in cases in which patients were still followed up in our institution. To
American Journal of Kidney Diseases, Vol 41, No 3 (March), 2003: pp 550-557
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
minimize possible bias, information obtained from patients and/or their parents was validated using data collected from available medical records. All patients gave their consent for this study.
Definitions Nephrotic syndrome (NS) is defined by proteinuria with protein of 0.04 g/h/m2 or greater in childhood and 3 g/d or greater in adulthood, hypoproteinemia (protein ⬍ 60 g/L), and hypoalbuminemia (albumin ⬍ 3 g/dL [30 g/L]). Remission of NS is defined by a reduction in urinary protein excretion rate to less than 4 mg/m2/h or by 0 to trace albuminuria on dipstick reading for at least 3 consecutive days. Relapse is defined by the reappearance of proteinuria (protein ⱖ 40 mg/h/m2 in childhood and ⱖ3 g/d in adulthood or dipstick ⫹⫹/⫹⫹⫹). SSNS diagnosis was made when remission of NS was obtained within 4 weeks after starting oral steroid therapy or eventually after the administration of intravenous steroids (discussed later). A patient was classified as a frequent relapser if he experienced more than three relapses during any 12-month period. Hypertension is defined in adults by a blood pressure of 140/90 mm Hg or greater. Osteoporosis is defined as a bone mineral density value, evaluated by absorptiometry of the lumbar spine and hip, less than ⫺2.5 SDs compared with the normal density for age and sex (z score). Weight excess is defined as a body mass index (BMI; weight [in kilograms]/length [square meters]) greater than 25 in men and greater than 24 in women. Obesity is defined as a BMI greater than 30. Short stature is defined as height less than 2 SDs compared with normal stature for age and sex.
Treatment Protocols The standard steroid regimen for the initial episode of NS was prednisone, 60 mg/m2 and no more than 80 mg/d for 1 month, followed by three methylprednisolone pulses (1,000 mg/1.73 m2 every other day) in case of persistent proteinuria. Thereafter, the patient was switched to alternate-day therapy. Total duration of therapy for the first attack varied from 1 to 4.5 months according to different local practices used in our institution or other institutions to which some patients were initially referred. Treatment of relapses consisted of administering daily prednisone, 40 to 60 mg/m2, until proteinuria had disappeared for 4 to 5 days. Thereafter, the regimen was switched to alternate days for 6 weeks. Dosage then was tapered progressively to 15 to 20 mg/m2 every other day according to the steroid threshold, ie, the dose at which the relapse had occurred. Treatment then was continued for 12 to 18 months. Levamisole was administered at a dosage of 2.5 mg/kg on alternate days, usually as a first-step change. Cytotoxic agents were administered to children requiring high-dose steroid therapy for maintenance of SSNS remission and presenting with severe steroid-related side effects (mainly statural growth retardation). Doses of cyclophosphamide, chlorambucil, and mechlorethamine were 2 mg/kg/d for 8 to 12 weeks, 0.2 to 0.3 mg/kg/d for 6 to 12 weeks, and 0.8 mg/kg (total dose) in two courses of four injections 1 month apart, respectively. Cyclosporine A therapy was started at a daily dose of 150 mg/m2.
551
Statistical Analysis All data are expressed as percentage or mean and 95% confidence interval. Chi-square and Student’s t-tests were used for percentage and mean comparisons. Multivariate analysis was performed using a logistic regression model with a backward procedure. Paired tests were used to compare mean numbers of NS relapses per patient per year during childhood, adolescence, and adulthood. The relationship between mean number of relapses per patient per year and age was studied using a linear marginal model in two conditions: adulthood nonrelapsers versus adulthood relapsers. This type of model takes into account repeated measures in each patient.10
RESULTS
Clinical Presentation in Childhood and Predictors of Adulthood Relapses One hundred seventeen patients were identified. Data regarding the outcome of NS in adulthood are available for 102 patients (87.2%). Forty-three patients (42.2%) experienced at least one relapse of NS in adulthood, whereas 59 patients (57.8%) did not experience relapse. Mean ages at time of last follow-up in relapsers and nonrelapsers were 25.9 years (95% confidence interval, 24.9 to 26.9) and 27.4 years (95% confidence interval, 26.8 to 28), respectively (P ⬍ 0.01). Characteristics during childhood of relapsers and nonrelapsers are listed in Table 1. The two groups did not differ with regard to sex ratio. Adulthood relapsers were younger at NS onset compared with nonrelapsers (mean age at onset, 4.1 years [95% confidence interval, 3.2 to 4.9] versus 5.8 years [95% confidence interval, 5 to 6.6]; P ⬍ 0.004), and NS started before the age of 6 years more frequently in the first group compared with the second group (75.4% versus 52.5%; P ⫽ 0.034). More patients in the relapser group had undergone at least one renal biopsy compared with nonrelapsers (73.8% versus 35.1%; P ⫽ 0.0003). Renal biopsies were performed (1) before the administration of cytotoxic drugs (local practice in our institution in the 1970s and early 1980s) and (2) before and during a treatment with cyclosporine. All biopsy specimens had shown minimal change disease lesions. NS relapsed more frequently during childhood in the adulthood relapser group compared with the nonrelapser group, exemplified by mean number of relapses per year (0.95 relapses/y [95% confidence interval, 0.84 to 1.06] versus 0.42 relapses/y [95%
552
FAKHOURI ET AL Table 1.
Characteristics During Childhood of Adulthood Relapsers and Nonrelapsers Adulthood Relapsers
No. of patients Women/men Birth weight (kg) Age at onset (ys) Disease onset ⬍ 6 y Kidney biopsy No. of relapses/patient/y No. of relapses in first 6 mon of disease Frequent relapsers Levamisole Cytotoxic agents Cyclosporine Age at last relapse (y)*
Adulthood Nonrelapsers
43/102 (42.2) 7/36 3.3 (3.05-3.55) 4.1 (3.2-4.9) 33/44 (75.0) 31/42 (73.8) 0.95 (0.84-1.06) 0.70 (0.5-0.91)
59/102 (57.8) 16/43 2.9 (2.64-3.16) 5.8 (5.0-6.6) 31/59 (52.5) 20/57 (35.1) 0.42 (0.33-0.51) 0.65 (0.46-0.84)
11/41 (26.8) 34/42 (80.9) 32/43 (74.4) 18/42 (42.9) 24 (22.9-25.01)
5/58 (8.6) 25/56 (44.6) 18/57 (31.6) 4/55 (7.3) 13.5 (13.32-13.68)
P
NS NS 0.004 0.034 0.0003 ⬍0.0001 NS 0.03 0.0006 ⬍0.0001 ⬍0.0001
NOTE. Values expressed as number/total number (percent) or mean (95% confidence interval). Abbreviation: NS, statistically not significant. *Age at last reported relapse for adulthood relapsers.
confidence interval, 5 to 6.6]; P ⬍ 0.0001) or number of frequent relapsers (26.8% versus 8.6%; P ⫽ 0.03). Conversely, the two groups did not differ regarding mean number of relapses in the first 6 months. Immunosuppressive or immunomodulator drugs were administered more frequently during childhood in the relapser group compared with the nonrelapser group (P ⬍ 0.006 for levamisole; P ⬍ 0.0001 for cytotoxic agents; P ⬍ 0.001 for cyclosporine; Table 1). Birth weight, available for 36 patients (15 patients, relapser group; 21 patients, nonrelapser group), did not differ between the two groups and did not correlate with total number of NS relapses in childhood or occurrence of adulthood SSNS relapses (data not shown). Multivariate analysis (logistic model) was performed using status of adulthood relapser as the variable to be explained and the following covariables: sex, age at onset, mean number of relapses per year during childhood, number of relapses in the first 6 months, and use of cytotoxic drugs, cyclosporine, and levamisole. Only mean number of relapses per year in childhood was predictive of SSNS relapses in adulthood (P ⫽ 0.0058). Relapses of NS in Adults Characteristics in adulthood of the relapser group are listed in Table 2. Evolution over time of mean number of relapses per patient per year for 36 adulthood
relapsers and 58 adulthood nonrelapsers is shown in Fig 1. An equation was calculated using a linear marginal model: y ⫽ 2.51 ⫺ 0.053x for adulthood relapsers and y ⫽ 2.09 ⫺ 0.11x for nonrelapsers. In both groups, number of relapses per patient per year decreased with age (P ⫽ 0.0003). Slopes of the two curves statistically differed (P ⫽ 0.008) in contrast to intercepts (P ⫽ 0.22). In adulthood nonrelapsers, mean age at the last SSNS relapse was 13.5 years (95% confidence interval, 13.32 to 13.68), and 36 of 56 patients (64.3%) experienced their last relapse between the age of 11 and 16 years (Fig 2). Mean age at the last reported relapse in adulthood relapsers was 24 years (95% confidence interval, 22.9 to 25.01; Table 1). Mean numbers of relapses per year for 36 adulthood relapsers during childTable 2.
Characteristics in Adulthood of Relapsing Patients
No. of patients Age at last follow-up (y) Length (cm) BMI (kg/m2) No. of relapses/y Levamisole Cytotoxic agents Cyclosporine
43 25.9 (24.9-26.9) 166 (163.7-168.3) 22.6 (21.3-23.9) 0.6 (0.44-0.76) 4/40 (10) 9/39 (23) 9/41 (22)
NOTE. Values expressed as number/total number (percent) or mean (95% confidence interval).
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
553
Fig 1. Evolution over time of mean number of NS relapse(s) per patient per year in 36 adulthood relapsers and 58 adulthood nonrelapsers. In both groups, number of relapses correlated with age (P ⴝ 0.0003). Slopes of the two curves statistically significantly differed (P ⴝ 0.008), whereas intercepts were not statistically significantly different (P ⴝ 0.22).
hood, puberty/adolescence, and adulthood are shown in Fig 3. At last follow-up in adulthood, 6 patients (14%) were experiencing an NS relapse. Twentyeight patients (65%) were still being administered steroids; three patients (7%), cytotoxic drugs; and three patients (7%), cyclosporine. Mean creatinine clearance was 112 mL/min (95% confidence interval, 13.32 to 13.68). NS or treatment-related complications in adulthood are listed in Table 3. One patient (2.3%) reached end-stage renal disease (ESRD) in the setting of a high-level corticodependent and cyclosporine-resistant NS. Steroid therapy was discontinued because of major side effects (osteoporosis and hypertension), and ESRD occurred 2
years after steroid withdrawal. Two renal biopsies performed in this patient in adulthood, 4 and 2 years before ESRD, showed minimal change disease in the absence of focal glomerulosclerosis lesions. All remaining patients retained normal renal function, including those administered cyclosporine. Late steroid resistance did not occur in any patient. Steroid-related complications occurred in 19 of 36 patients (44.2%): osteoporosis (7 of 11 patients), weight excess (8 of 43 patients), short stature (6 of 36 patients), hypertension (3 of 43 patients), myocardial infarction (1 of 43 patients), non–insulin-dependent diabetes mellitus (1 of 43 patients), and psychiatric disorders (1 of 43 patients). NS-related thromboembolic complications occurred in 2 of 43 pa-
Fig 2. Age at last NS relapse in 56 adulthood nonrelapsers.
554
FAKHOURI ET AL
Fig 3. Mean number of NS relapse(s) per patient per year during childhood, puberty/late adolescence, and adulthood in 36 adulthood relapsers.
tients (4.7%): pulmonary embolism in 1 patient and cerebral thrombophlebitis in 1 patient. DISCUSSION
Few data regarding the outcome in adulthood of childhood SSNS have been previously published. In the largest series of childhood SSNS, the International Study of Kidney Disease in Children (ISKDC),2 all information regarding the adulthood outcome of this disease is confined to the statement that “5 to 10% of patients. . .continued to relapse at the last follow-up,” presumably during late adolescence or early adulthood. Similarly, in two other studies of long-term outcome of childhood SSNS published in The Lancet in the 1980s,8,9 data regarding adulthood outcome of the disease are scarce. In the first report by Trompeter et al,8 which Table 3. NS and/or Treatment-Related Complications During Adulthood in Relapsing Patients ESRD Short stature Weight excess Obesity Steroid-induced osteoporosis Hypertension Myocardial infarction/heart failure Cerebral thrombophlebitis Pulmonary embolism Non–insulin-dependent diabetes mellitus Psychiatric disorders
1/43 (2.3) 6/36 (16) 8/43 (19) 2/43 (4.6) 7/11 (63) 3/43 (6.9) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3)
NOTE. Values expressed as number of patients/total patients (percent).
included 10 patients (of 183 children) still experiencing relapses in adulthood, the number of patients followed up in adulthood is not specified and the follow-up period did not extend beyond the age of 20 years. The second report by Lewis et al9 included only 26 patients older than 20 years at last follow-up, of whom 5 patients (19.2%) experienced a relapse in adulthood. Duration of follow-up in adulthood was not specified. Similarly, other studies addressing this issue included a small number of patients and/or reported a short duration of follow-up in adulthood.11-13 Moreover, most of these studies were conducted in the precyclosporine era. Our study documents for the first time the long-term outcome in adulthood of a large cohort of patients diagnosed with an SSNS in childhood. However, it carries the limitations and possible bias of all retrospective studies, particularly in data collection. Relapses in adulthood occurred in more than 40% of our patients. This incidence is greater than that suggested by previous reports.2,8,9 Several reasons may explain this difference. First, a larger number of patients followed up in adulthood was included in our survey compared with previous studies. Second, because the Paediatric Nephrology Department at the Hoˆ pital NeckerEnfants Malades is a second-referral center, our cohort probably included severe cases of childhood SSNS. The percentage of patients experiencing frequent relapses in our series (16%) was relatively low compared with the ISKDC study (34%).2 However, the prolonged steroid regimen
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
555
(12 to 18 months) used in our pediatric institution for the treatment of SSNS relapses clearly underestimates the number of frequent relapses. Conversely, the use of cytotoxic agents, another reflection of disease severity, was less frequent in our study (49%) compared with the report by Lewis et al9 (72%), although cytotoxic agent indications may vary from one center to another. However, only 15% of patients (16 of 102 patients), all adulthood nonrelapsers, experienced two or fewer relapses in all (data not shown). Finally, the most interesting aspects of SSNS, such as treatment regimen, are encountered almost exclusively in severe forms of the disease, ie, frequently relapsing and steroid-dependent patients. Thus, our cohort is representative of the most challenging cases of (mainly steroiddependent) patients with SSNS requiring prolonged follow-up in childhood and with an increasing frequency in adulthood medical care centers. Not surprisingly, frequently relapsing disease in childhood requiring the use of immunosuppressive and/or immunomodulator drugs was predictive of SSNS relapses in adulthood. In univariate, but not multivariate, analysis, young age at onset, mainly younger than 6 years, also correlated with the occurrence of adulthood relapses. This feature stands in contrast to data reported by Lewis et al,9 but is in accordance with observations of Trompeter et al.8 Conversely, the number of relapses in the first 6 months of disease, a powerful predictor of disease severity in childhood according to the ISKDC,2 failed to distinguish adulthood relapsers from nonrelapsers. Previous reports of small series of patients with SSNS had documented more severe disease (ie, frequent relapses, use of cytotoxic agents) in children with intrauterine growth retardation.14,15 In our series, birth weight did not correlate with number of relapses in childhood or occurrence of adulthood relapses. However, data were available for approximately one third of patients, of whom only two patients (2%) had intrauterine growth retardation. In adulthood relapsers and nonrelapsers, disease activity steadily decreased over time (Fig 1). The two groups did not differ regarding the initial relapse rate, shown in Fig 1 and exemplified by number of relapses in the first 6 months. However, the decline in disease activity was
more rapid in adulthood nonrelapsers compared with relapsers, allowing an end to the disease before adulthood. It is noteworthy that in almost two thirds of adulthood nonrelapsers, NS subsided during puberty (Fig 2), probably from the steady decrease in disease activity in the absence of a clear impact of puberty on the SSNS relapse rate in adulthood nonrelapsers, as well as relapsers (Fig 1). In adulthood relapsers, mean number of relapses per year in adulthood was 0.56 compared with 1.07 and 0.79 in childhood and puberty/ adolescence, respectively (Fig 3). However, this less severe presentation of SSNS in adulthood is hampered by the occurrence of serious NSrelated complications, mainly thromboembolic, in some patients (Table 3). In addition, relapse rate increased during adulthood compared with childhood in 10 of 36 patients (27.8%) (data not shown), and almost two thirds of adulthood relapsers were still administered steroids at last follow-up. Conversely, the use of immunosuppressive or immunomodulator drugs was fairly rare because of two factors: the decreasing frequency of relapses and that three quarters of adulthood relapsers had already been administered cytotoxic agents during childhood, the toxic effects of which are cumulative. Overall, renal outcome was favorable. Only 1 patient progressed to ESRD in the setting of a high-level steroid-dependent and cyclosporineresistant NS. ESRD occurred 2 years after the discontinuation of steroid therapy. It is worth noting that in adulthood, this patient underwent two renal biopsies that failed to show focal glomerulosclerotic lesions. Such an unfavorable outcome of typical minimal change disease has been previously, although very rarely, reported.1,16 All remaining patients retained normal renal function, including those who had been administered cyclosporine. However, long-term renal tolerance of cyclosporine remains a matter of concern regarding renal function in adulthood relapsers. In idiopathic NS, cyclosporine may prove detrimental to renal function in patients with focal segmental glomerulosclerosis, especially when renal failure and interstitial fibrosis preexist to cyclosporine use.17 Although cyclosporine may be less nephrotoxic in minimal change disease,1 its long-term impact on renal function after sev-
556
FAKHOURI ET AL
eral years of administration remains to be evaluated. Long-term steroid therapy started during childhood is associated with a number of significant side effects, mainly growth retardation and/or obesity, osteoporosis, and cardiovascular events. Such side effects were noted in 44.2% of our patients. Short stature and obesity were fairly rare in our adulthood relapsers, in accordance with previous reports.2,8,18 The occurrence of growth retardation and/or weight excess during childhood usually prompts pediatricians to use such steroid-sparing agents as levamisole, cytotoxic drugs, or cyclosporine. Steroid-induced osteoporosis was noted in 7 of 11 evaluated patients. Bone mineral density was not monitored during childhood. Bisphosphonates have been recommended in the prevention and treatment of corticosteroid-induced osteoporosis in adults,19 although their impact on risk for bone fractures has not been clearly established, except perhaps in postmenopausal women.20 Conversely, there is no clear consensus regarding their use in children. The high incidence of osteoporosis in our adult patients clearly indicates the need for studies regarding the prevention of steroidinduced osteoporosis in children. The cardiovascular impact of long-term steroid therapy may have been underestimated in our retrospective study. This impact will probably become more apparent as adulthood relapsers enter their 40s or 50s. Unfortunately, there are no available data concerning effects of cytotoxic drugs on fertility in our patients. In conclusion, almost half the patients with childhood SSNS included in our series experienced relapses in adulthood. The second-referral status of our pediatric institution may have contributed to this high frequency of adulthood relapses. During the next several decades, additional studies are required to assess delayedonset complications of long-term steroid therapy, as well as long-term consequences of cyclosporine, which has been used as a substitute for immunosuppressors in children with high-level steroid dependency or frequent relapses. ACKNOWLEDGMENT The authors thank all patients, their parents, and attending physicians for providing information, and Victoria Nagel-
Hauzy and Claude de la Potterie for invaluable assistance in manuscript preparation.
REFERENCES 1. Broyer M, Meyrier A, Niaudet P, Habib R: Minimal changes and focal segmental sclerosis, in Davison AM, Cameron JS, Gru¨ nfeld JP, Kerr DN, Ritz E, Winearls EG (eds): Oxford Textbook of Clinical Nephrology (ed 2). Oxford, NY: Oxford University Press, 1998, pp 494-535 2. Tarshish P, Tobin JN, Bernstein J, Edelmann CM Jr: 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 8:769-776, 1997 3. A Report of the International Study of Kidney Disease in Children: Primary nephrotic syndrome in children: Clinical significance of histopathologic variants of minimal change and of diffuse mesangial hypercellularity. Kidney Int 20:765771, 1981 4. Report of the International Study of Kidney Disease in Children: Prospective, controlled trial of cyclophosphamide therapy in children with nephrotic syndrome. Lancet 2:423427, 1974 5. Grupe WE, Makker SP, Ingelfinger JR: Chlorambucil treatment of frequently relapsing nephrotic syndrome. N Engl J Med 295:746-749, 1976 6. Niaudet P: The French Society of Paediatric Nephrology: Comparison of cyclosporin and chlorambucil in the treatment of steroid-dependent idiopathic nephrotic syndrome: A multicentre randomized controlled trial. Pediatr Nephrol 6:1-3, 1992 7. British Association for Paediatric Nephrology: Levamisole for corticosteroid-dependent nephrotic syndrome in childhood. Lancet 337:1555-1557, 1991 8. Trompeter RS, Lloyd BW, Hicks J, White RH, Cameron JS: Long-term outcome for children with minimalchange nephrotic syndrome. Lancet 1:368-370, 1985 9. Lewis MA, Baildom EM, Davis N, Houston IB, Postlethwaite RJ: Nephrotic syndrome: From toddlers to twenties. Lancet 1:255-259, 1989 10. Liang K-Y, Zeger S: Longitudinal data analysis using generalized linear model. Biometrika 73:13-22, 1986 11. Siegel NJ, Goldberg B, Krassner LS, Hayslett JP: Long-term follow-up of children with steroid-responsive nephrotic syndrome. J Pediatr 81:251-258, 1972 12. Schwartz MW, Schwartz GJ, Cornfeld D: A 16-year follow-up study of 163 children with nephrotic syndrome. Pediatrics 54:547-552, 1974 13. Berns JS, Gaudio KM, Krassner LS, et al: Steroidresponsive nephrotic syndrome of childhood: A long-term study of clinical course, histopathology, efficacy of cyclophosphamide therapy, and effects on growth. Am J Kidney Dis 9:108-114, 1987 14. Zidar N, Avgustin Cavic M, Kenda RB, Ferluga D: Unfavorable course of minimal change nephrotic syndrome in children with intrauterine growth retardation. Kidney Int 54:1320-1323, 1998 15. Sheu JN, Chen JH: Minimal change nephrotic syndrome in children with intrauterine growth retardation. Am J Kidney Dis 37:909-914, 2001 16. Hayslett JP, Krassner LS, Bensch KG, Kashgarian M,
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
557
Epstein FH: Progression of “lipoid nephrosis” to renal insufficiency. N Engl J Med 281:181-187, 1969 17. Meyrier A, Noe¨ l LH, Auriche P, Callard P. Collaborative Group of the Socie´ te´ de Ne´ phrologie: Long-term renal tolerance of cyclosporin A treatment in adult idiopathic nephrotic syndrome. Kidney Int 45:1446-1456, 1994 18. Foote KD, Brocklebank JT, Meadow SR: Height attainment in children with steroid-responsive nephrotic syndrome. Lancet 2:917-919, 1985
19. American College of Rheumatology Ad Hoc Committee on Glucocorticoid-Induced Osteoporosis: Recommendations for the prevention and treatment of glucocorticoidinduced osteoporosis: 2001 Update. Arthritis Rheum 44: 1496-1503, 2001 20. McClung MR, Geusens P, Miller PD, et al: Hip Intervention Program Study Group. Effect of risedronate on the risk of hip fracture in elderly women. N Engl J Med 344:333-340, 2001
S
TEROID-SENSITIVE nephrotic syndrome (SSNS) is the most frequent glomerular disease encountered during childhood.1 Pathophysiological characteristics of SSNS, most often related to minimal change disease, remain elusive. Previous studies have extensively assessed the clinical presentation, treatment, and outcome of this disease during childhood.2-5 SSNS relapses in up to 70% of children, with most of these patients frequent relapsers or steroid dependent.2 Immunosuppressive or immunomodulator drugs, including cyclophosphamide,4 chlorambucil,5 cyclosporine A,6 or levamisole,7 have proved effective as steroid-sparing agents From the Departments of Nephrology, Pediatric Nephrology, and Biostatistics, Assistance Publique-Hoˆpitaux de Paris, Hoˆpital Necker-Enfants Malades, Paris, France. Received June 22, 2002; accepted in revised form October 10, 2002. Address reprint requests to Fadi Fakhouri, MD, Service de Ne´phrologie, Hoˆpital Necker, 149 rue de Se`vres, 75743 Paris Cedex 15, France. E-mail: [email protected] © 2003 by the National Kidney Foundation, Inc. 0272-6386/03/4103-0003$30.00/0 doi:10.1053/ajkd.2003.50116 550
in frequently relapsing patients. Renal prognosis is usually favorable in these children despite their frequent relapses. Few data regarding the outcome in adulthood of childhood SSNS are available.8,9 Thus, the risk for relapse and incidence of treatment-related complications in adulthood remain largely unknown. This report is the first extensive study regarding the outcome in adulthood of a large group of patients diagnosed with SSNS during childhood. PATIENTS AND METHODS
Patients Using a computerized file, we identified all children: (1) with an SSNS, (2) born between 1970 and 1975 (thus aged ⬎ 25 years at the time of the study), and (3) who had been followed up (as inpatients or outpatients) in the Department of Paediatric Nephrology at the Hoˆpital NeckerEnfants Malades (Paris, France). Pediatric medical records of these patients were reviewed to collect relevant data. Information regarding the outcome of patients in adulthood (ⱖ18 years) was obtained from mailed questionnaires and phone calls to the patients and/or their parents, from adult nephrologists or general practitioners, and from medical records at Hoˆpital Necker-Enfants Malades in cases in which patients were still followed up in our institution. To
American Journal of Kidney Diseases, Vol 41, No 3 (March), 2003: pp 550-557
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
minimize possible bias, information obtained from patients and/or their parents was validated using data collected from available medical records. All patients gave their consent for this study.
Definitions Nephrotic syndrome (NS) is defined by proteinuria with protein of 0.04 g/h/m2 or greater in childhood and 3 g/d or greater in adulthood, hypoproteinemia (protein ⬍ 60 g/L), and hypoalbuminemia (albumin ⬍ 3 g/dL [30 g/L]). Remission of NS is defined by a reduction in urinary protein excretion rate to less than 4 mg/m2/h or by 0 to trace albuminuria on dipstick reading for at least 3 consecutive days. Relapse is defined by the reappearance of proteinuria (protein ⱖ 40 mg/h/m2 in childhood and ⱖ3 g/d in adulthood or dipstick ⫹⫹/⫹⫹⫹). SSNS diagnosis was made when remission of NS was obtained within 4 weeks after starting oral steroid therapy or eventually after the administration of intravenous steroids (discussed later). A patient was classified as a frequent relapser if he experienced more than three relapses during any 12-month period. Hypertension is defined in adults by a blood pressure of 140/90 mm Hg or greater. Osteoporosis is defined as a bone mineral density value, evaluated by absorptiometry of the lumbar spine and hip, less than ⫺2.5 SDs compared with the normal density for age and sex (z score). Weight excess is defined as a body mass index (BMI; weight [in kilograms]/length [square meters]) greater than 25 in men and greater than 24 in women. Obesity is defined as a BMI greater than 30. Short stature is defined as height less than 2 SDs compared with normal stature for age and sex.
Treatment Protocols The standard steroid regimen for the initial episode of NS was prednisone, 60 mg/m2 and no more than 80 mg/d for 1 month, followed by three methylprednisolone pulses (1,000 mg/1.73 m2 every other day) in case of persistent proteinuria. Thereafter, the patient was switched to alternate-day therapy. Total duration of therapy for the first attack varied from 1 to 4.5 months according to different local practices used in our institution or other institutions to which some patients were initially referred. Treatment of relapses consisted of administering daily prednisone, 40 to 60 mg/m2, until proteinuria had disappeared for 4 to 5 days. Thereafter, the regimen was switched to alternate days for 6 weeks. Dosage then was tapered progressively to 15 to 20 mg/m2 every other day according to the steroid threshold, ie, the dose at which the relapse had occurred. Treatment then was continued for 12 to 18 months. Levamisole was administered at a dosage of 2.5 mg/kg on alternate days, usually as a first-step change. Cytotoxic agents were administered to children requiring high-dose steroid therapy for maintenance of SSNS remission and presenting with severe steroid-related side effects (mainly statural growth retardation). Doses of cyclophosphamide, chlorambucil, and mechlorethamine were 2 mg/kg/d for 8 to 12 weeks, 0.2 to 0.3 mg/kg/d for 6 to 12 weeks, and 0.8 mg/kg (total dose) in two courses of four injections 1 month apart, respectively. Cyclosporine A therapy was started at a daily dose of 150 mg/m2.
551
Statistical Analysis All data are expressed as percentage or mean and 95% confidence interval. Chi-square and Student’s t-tests were used for percentage and mean comparisons. Multivariate analysis was performed using a logistic regression model with a backward procedure. Paired tests were used to compare mean numbers of NS relapses per patient per year during childhood, adolescence, and adulthood. The relationship between mean number of relapses per patient per year and age was studied using a linear marginal model in two conditions: adulthood nonrelapsers versus adulthood relapsers. This type of model takes into account repeated measures in each patient.10
RESULTS
Clinical Presentation in Childhood and Predictors of Adulthood Relapses One hundred seventeen patients were identified. Data regarding the outcome of NS in adulthood are available for 102 patients (87.2%). Forty-three patients (42.2%) experienced at least one relapse of NS in adulthood, whereas 59 patients (57.8%) did not experience relapse. Mean ages at time of last follow-up in relapsers and nonrelapsers were 25.9 years (95% confidence interval, 24.9 to 26.9) and 27.4 years (95% confidence interval, 26.8 to 28), respectively (P ⬍ 0.01). Characteristics during childhood of relapsers and nonrelapsers are listed in Table 1. The two groups did not differ with regard to sex ratio. Adulthood relapsers were younger at NS onset compared with nonrelapsers (mean age at onset, 4.1 years [95% confidence interval, 3.2 to 4.9] versus 5.8 years [95% confidence interval, 5 to 6.6]; P ⬍ 0.004), and NS started before the age of 6 years more frequently in the first group compared with the second group (75.4% versus 52.5%; P ⫽ 0.034). More patients in the relapser group had undergone at least one renal biopsy compared with nonrelapsers (73.8% versus 35.1%; P ⫽ 0.0003). Renal biopsies were performed (1) before the administration of cytotoxic drugs (local practice in our institution in the 1970s and early 1980s) and (2) before and during a treatment with cyclosporine. All biopsy specimens had shown minimal change disease lesions. NS relapsed more frequently during childhood in the adulthood relapser group compared with the nonrelapser group, exemplified by mean number of relapses per year (0.95 relapses/y [95% confidence interval, 0.84 to 1.06] versus 0.42 relapses/y [95%
552
FAKHOURI ET AL Table 1.
Characteristics During Childhood of Adulthood Relapsers and Nonrelapsers Adulthood Relapsers
No. of patients Women/men Birth weight (kg) Age at onset (ys) Disease onset ⬍ 6 y Kidney biopsy No. of relapses/patient/y No. of relapses in first 6 mon of disease Frequent relapsers Levamisole Cytotoxic agents Cyclosporine Age at last relapse (y)*
Adulthood Nonrelapsers
43/102 (42.2) 7/36 3.3 (3.05-3.55) 4.1 (3.2-4.9) 33/44 (75.0) 31/42 (73.8) 0.95 (0.84-1.06) 0.70 (0.5-0.91)
59/102 (57.8) 16/43 2.9 (2.64-3.16) 5.8 (5.0-6.6) 31/59 (52.5) 20/57 (35.1) 0.42 (0.33-0.51) 0.65 (0.46-0.84)
11/41 (26.8) 34/42 (80.9) 32/43 (74.4) 18/42 (42.9) 24 (22.9-25.01)
5/58 (8.6) 25/56 (44.6) 18/57 (31.6) 4/55 (7.3) 13.5 (13.32-13.68)
P
NS NS 0.004 0.034 0.0003 ⬍0.0001 NS 0.03 0.0006 ⬍0.0001 ⬍0.0001
NOTE. Values expressed as number/total number (percent) or mean (95% confidence interval). Abbreviation: NS, statistically not significant. *Age at last reported relapse for adulthood relapsers.
confidence interval, 5 to 6.6]; P ⬍ 0.0001) or number of frequent relapsers (26.8% versus 8.6%; P ⫽ 0.03). Conversely, the two groups did not differ regarding mean number of relapses in the first 6 months. Immunosuppressive or immunomodulator drugs were administered more frequently during childhood in the relapser group compared with the nonrelapser group (P ⬍ 0.006 for levamisole; P ⬍ 0.0001 for cytotoxic agents; P ⬍ 0.001 for cyclosporine; Table 1). Birth weight, available for 36 patients (15 patients, relapser group; 21 patients, nonrelapser group), did not differ between the two groups and did not correlate with total number of NS relapses in childhood or occurrence of adulthood SSNS relapses (data not shown). Multivariate analysis (logistic model) was performed using status of adulthood relapser as the variable to be explained and the following covariables: sex, age at onset, mean number of relapses per year during childhood, number of relapses in the first 6 months, and use of cytotoxic drugs, cyclosporine, and levamisole. Only mean number of relapses per year in childhood was predictive of SSNS relapses in adulthood (P ⫽ 0.0058). Relapses of NS in Adults Characteristics in adulthood of the relapser group are listed in Table 2. Evolution over time of mean number of relapses per patient per year for 36 adulthood
relapsers and 58 adulthood nonrelapsers is shown in Fig 1. An equation was calculated using a linear marginal model: y ⫽ 2.51 ⫺ 0.053x for adulthood relapsers and y ⫽ 2.09 ⫺ 0.11x for nonrelapsers. In both groups, number of relapses per patient per year decreased with age (P ⫽ 0.0003). Slopes of the two curves statistically differed (P ⫽ 0.008) in contrast to intercepts (P ⫽ 0.22). In adulthood nonrelapsers, mean age at the last SSNS relapse was 13.5 years (95% confidence interval, 13.32 to 13.68), and 36 of 56 patients (64.3%) experienced their last relapse between the age of 11 and 16 years (Fig 2). Mean age at the last reported relapse in adulthood relapsers was 24 years (95% confidence interval, 22.9 to 25.01; Table 1). Mean numbers of relapses per year for 36 adulthood relapsers during childTable 2.
Characteristics in Adulthood of Relapsing Patients
No. of patients Age at last follow-up (y) Length (cm) BMI (kg/m2) No. of relapses/y Levamisole Cytotoxic agents Cyclosporine
43 25.9 (24.9-26.9) 166 (163.7-168.3) 22.6 (21.3-23.9) 0.6 (0.44-0.76) 4/40 (10) 9/39 (23) 9/41 (22)
NOTE. Values expressed as number/total number (percent) or mean (95% confidence interval).
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
553
Fig 1. Evolution over time of mean number of NS relapse(s) per patient per year in 36 adulthood relapsers and 58 adulthood nonrelapsers. In both groups, number of relapses correlated with age (P ⴝ 0.0003). Slopes of the two curves statistically significantly differed (P ⴝ 0.008), whereas intercepts were not statistically significantly different (P ⴝ 0.22).
hood, puberty/adolescence, and adulthood are shown in Fig 3. At last follow-up in adulthood, 6 patients (14%) were experiencing an NS relapse. Twentyeight patients (65%) were still being administered steroids; three patients (7%), cytotoxic drugs; and three patients (7%), cyclosporine. Mean creatinine clearance was 112 mL/min (95% confidence interval, 13.32 to 13.68). NS or treatment-related complications in adulthood are listed in Table 3. One patient (2.3%) reached end-stage renal disease (ESRD) in the setting of a high-level corticodependent and cyclosporine-resistant NS. Steroid therapy was discontinued because of major side effects (osteoporosis and hypertension), and ESRD occurred 2
years after steroid withdrawal. Two renal biopsies performed in this patient in adulthood, 4 and 2 years before ESRD, showed minimal change disease in the absence of focal glomerulosclerosis lesions. All remaining patients retained normal renal function, including those administered cyclosporine. Late steroid resistance did not occur in any patient. Steroid-related complications occurred in 19 of 36 patients (44.2%): osteoporosis (7 of 11 patients), weight excess (8 of 43 patients), short stature (6 of 36 patients), hypertension (3 of 43 patients), myocardial infarction (1 of 43 patients), non–insulin-dependent diabetes mellitus (1 of 43 patients), and psychiatric disorders (1 of 43 patients). NS-related thromboembolic complications occurred in 2 of 43 pa-
Fig 2. Age at last NS relapse in 56 adulthood nonrelapsers.
554
FAKHOURI ET AL
Fig 3. Mean number of NS relapse(s) per patient per year during childhood, puberty/late adolescence, and adulthood in 36 adulthood relapsers.
tients (4.7%): pulmonary embolism in 1 patient and cerebral thrombophlebitis in 1 patient. DISCUSSION
Few data regarding the outcome in adulthood of childhood SSNS have been previously published. In the largest series of childhood SSNS, the International Study of Kidney Disease in Children (ISKDC),2 all information regarding the adulthood outcome of this disease is confined to the statement that “5 to 10% of patients. . .continued to relapse at the last follow-up,” presumably during late adolescence or early adulthood. Similarly, in two other studies of long-term outcome of childhood SSNS published in The Lancet in the 1980s,8,9 data regarding adulthood outcome of the disease are scarce. In the first report by Trompeter et al,8 which Table 3. NS and/or Treatment-Related Complications During Adulthood in Relapsing Patients ESRD Short stature Weight excess Obesity Steroid-induced osteoporosis Hypertension Myocardial infarction/heart failure Cerebral thrombophlebitis Pulmonary embolism Non–insulin-dependent diabetes mellitus Psychiatric disorders
1/43 (2.3) 6/36 (16) 8/43 (19) 2/43 (4.6) 7/11 (63) 3/43 (6.9) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3) 1/43 (2.3)
NOTE. Values expressed as number of patients/total patients (percent).
included 10 patients (of 183 children) still experiencing relapses in adulthood, the number of patients followed up in adulthood is not specified and the follow-up period did not extend beyond the age of 20 years. The second report by Lewis et al9 included only 26 patients older than 20 years at last follow-up, of whom 5 patients (19.2%) experienced a relapse in adulthood. Duration of follow-up in adulthood was not specified. Similarly, other studies addressing this issue included a small number of patients and/or reported a short duration of follow-up in adulthood.11-13 Moreover, most of these studies were conducted in the precyclosporine era. Our study documents for the first time the long-term outcome in adulthood of a large cohort of patients diagnosed with an SSNS in childhood. However, it carries the limitations and possible bias of all retrospective studies, particularly in data collection. Relapses in adulthood occurred in more than 40% of our patients. This incidence is greater than that suggested by previous reports.2,8,9 Several reasons may explain this difference. First, a larger number of patients followed up in adulthood was included in our survey compared with previous studies. Second, because the Paediatric Nephrology Department at the Hoˆ pital NeckerEnfants Malades is a second-referral center, our cohort probably included severe cases of childhood SSNS. The percentage of patients experiencing frequent relapses in our series (16%) was relatively low compared with the ISKDC study (34%).2 However, the prolonged steroid regimen
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
555
(12 to 18 months) used in our pediatric institution for the treatment of SSNS relapses clearly underestimates the number of frequent relapses. Conversely, the use of cytotoxic agents, another reflection of disease severity, was less frequent in our study (49%) compared with the report by Lewis et al9 (72%), although cytotoxic agent indications may vary from one center to another. However, only 15% of patients (16 of 102 patients), all adulthood nonrelapsers, experienced two or fewer relapses in all (data not shown). Finally, the most interesting aspects of SSNS, such as treatment regimen, are encountered almost exclusively in severe forms of the disease, ie, frequently relapsing and steroid-dependent patients. Thus, our cohort is representative of the most challenging cases of (mainly steroiddependent) patients with SSNS requiring prolonged follow-up in childhood and with an increasing frequency in adulthood medical care centers. Not surprisingly, frequently relapsing disease in childhood requiring the use of immunosuppressive and/or immunomodulator drugs was predictive of SSNS relapses in adulthood. In univariate, but not multivariate, analysis, young age at onset, mainly younger than 6 years, also correlated with the occurrence of adulthood relapses. This feature stands in contrast to data reported by Lewis et al,9 but is in accordance with observations of Trompeter et al.8 Conversely, the number of relapses in the first 6 months of disease, a powerful predictor of disease severity in childhood according to the ISKDC,2 failed to distinguish adulthood relapsers from nonrelapsers. Previous reports of small series of patients with SSNS had documented more severe disease (ie, frequent relapses, use of cytotoxic agents) in children with intrauterine growth retardation.14,15 In our series, birth weight did not correlate with number of relapses in childhood or occurrence of adulthood relapses. However, data were available for approximately one third of patients, of whom only two patients (2%) had intrauterine growth retardation. In adulthood relapsers and nonrelapsers, disease activity steadily decreased over time (Fig 1). The two groups did not differ regarding the initial relapse rate, shown in Fig 1 and exemplified by number of relapses in the first 6 months. However, the decline in disease activity was
more rapid in adulthood nonrelapsers compared with relapsers, allowing an end to the disease before adulthood. It is noteworthy that in almost two thirds of adulthood nonrelapsers, NS subsided during puberty (Fig 2), probably from the steady decrease in disease activity in the absence of a clear impact of puberty on the SSNS relapse rate in adulthood nonrelapsers, as well as relapsers (Fig 1). In adulthood relapsers, mean number of relapses per year in adulthood was 0.56 compared with 1.07 and 0.79 in childhood and puberty/ adolescence, respectively (Fig 3). However, this less severe presentation of SSNS in adulthood is hampered by the occurrence of serious NSrelated complications, mainly thromboembolic, in some patients (Table 3). In addition, relapse rate increased during adulthood compared with childhood in 10 of 36 patients (27.8%) (data not shown), and almost two thirds of adulthood relapsers were still administered steroids at last follow-up. Conversely, the use of immunosuppressive or immunomodulator drugs was fairly rare because of two factors: the decreasing frequency of relapses and that three quarters of adulthood relapsers had already been administered cytotoxic agents during childhood, the toxic effects of which are cumulative. Overall, renal outcome was favorable. Only 1 patient progressed to ESRD in the setting of a high-level steroid-dependent and cyclosporineresistant NS. ESRD occurred 2 years after the discontinuation of steroid therapy. It is worth noting that in adulthood, this patient underwent two renal biopsies that failed to show focal glomerulosclerotic lesions. Such an unfavorable outcome of typical minimal change disease has been previously, although very rarely, reported.1,16 All remaining patients retained normal renal function, including those who had been administered cyclosporine. However, long-term renal tolerance of cyclosporine remains a matter of concern regarding renal function in adulthood relapsers. In idiopathic NS, cyclosporine may prove detrimental to renal function in patients with focal segmental glomerulosclerosis, especially when renal failure and interstitial fibrosis preexist to cyclosporine use.17 Although cyclosporine may be less nephrotoxic in minimal change disease,1 its long-term impact on renal function after sev-
556
FAKHOURI ET AL
eral years of administration remains to be evaluated. Long-term steroid therapy started during childhood is associated with a number of significant side effects, mainly growth retardation and/or obesity, osteoporosis, and cardiovascular events. Such side effects were noted in 44.2% of our patients. Short stature and obesity were fairly rare in our adulthood relapsers, in accordance with previous reports.2,8,18 The occurrence of growth retardation and/or weight excess during childhood usually prompts pediatricians to use such steroid-sparing agents as levamisole, cytotoxic drugs, or cyclosporine. Steroid-induced osteoporosis was noted in 7 of 11 evaluated patients. Bone mineral density was not monitored during childhood. Bisphosphonates have been recommended in the prevention and treatment of corticosteroid-induced osteoporosis in adults,19 although their impact on risk for bone fractures has not been clearly established, except perhaps in postmenopausal women.20 Conversely, there is no clear consensus regarding their use in children. The high incidence of osteoporosis in our adult patients clearly indicates the need for studies regarding the prevention of steroidinduced osteoporosis in children. The cardiovascular impact of long-term steroid therapy may have been underestimated in our retrospective study. This impact will probably become more apparent as adulthood relapsers enter their 40s or 50s. Unfortunately, there are no available data concerning effects of cytotoxic drugs on fertility in our patients. In conclusion, almost half the patients with childhood SSNS included in our series experienced relapses in adulthood. The second-referral status of our pediatric institution may have contributed to this high frequency of adulthood relapses. During the next several decades, additional studies are required to assess delayedonset complications of long-term steroid therapy, as well as long-term consequences of cyclosporine, which has been used as a substitute for immunosuppressors in children with high-level steroid dependency or frequent relapses. ACKNOWLEDGMENT The authors thank all patients, their parents, and attending physicians for providing information, and Victoria Nagel-
Hauzy and Claude de la Potterie for invaluable assistance in manuscript preparation.
REFERENCES 1. Broyer M, Meyrier A, Niaudet P, Habib R: Minimal changes and focal segmental sclerosis, in Davison AM, Cameron JS, Gru¨ nfeld JP, Kerr DN, Ritz E, Winearls EG (eds): Oxford Textbook of Clinical Nephrology (ed 2). Oxford, NY: Oxford University Press, 1998, pp 494-535 2. Tarshish P, Tobin JN, Bernstein J, Edelmann CM Jr: 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 8:769-776, 1997 3. A Report of the International Study of Kidney Disease in Children: Primary nephrotic syndrome in children: Clinical significance of histopathologic variants of minimal change and of diffuse mesangial hypercellularity. Kidney Int 20:765771, 1981 4. Report of the International Study of Kidney Disease in Children: Prospective, controlled trial of cyclophosphamide therapy in children with nephrotic syndrome. Lancet 2:423427, 1974 5. Grupe WE, Makker SP, Ingelfinger JR: Chlorambucil treatment of frequently relapsing nephrotic syndrome. N Engl J Med 295:746-749, 1976 6. Niaudet P: The French Society of Paediatric Nephrology: Comparison of cyclosporin and chlorambucil in the treatment of steroid-dependent idiopathic nephrotic syndrome: A multicentre randomized controlled trial. Pediatr Nephrol 6:1-3, 1992 7. British Association for Paediatric Nephrology: Levamisole for corticosteroid-dependent nephrotic syndrome in childhood. Lancet 337:1555-1557, 1991 8. Trompeter RS, Lloyd BW, Hicks J, White RH, Cameron JS: Long-term outcome for children with minimalchange nephrotic syndrome. Lancet 1:368-370, 1985 9. Lewis MA, Baildom EM, Davis N, Houston IB, Postlethwaite RJ: Nephrotic syndrome: From toddlers to twenties. Lancet 1:255-259, 1989 10. Liang K-Y, Zeger S: Longitudinal data analysis using generalized linear model. Biometrika 73:13-22, 1986 11. Siegel NJ, Goldberg B, Krassner LS, Hayslett JP: Long-term follow-up of children with steroid-responsive nephrotic syndrome. J Pediatr 81:251-258, 1972 12. Schwartz MW, Schwartz GJ, Cornfeld D: A 16-year follow-up study of 163 children with nephrotic syndrome. Pediatrics 54:547-552, 1974 13. Berns JS, Gaudio KM, Krassner LS, et al: Steroidresponsive nephrotic syndrome of childhood: A long-term study of clinical course, histopathology, efficacy of cyclophosphamide therapy, and effects on growth. Am J Kidney Dis 9:108-114, 1987 14. Zidar N, Avgustin Cavic M, Kenda RB, Ferluga D: Unfavorable course of minimal change nephrotic syndrome in children with intrauterine growth retardation. Kidney Int 54:1320-1323, 1998 15. Sheu JN, Chen JH: Minimal change nephrotic syndrome in children with intrauterine growth retardation. Am J Kidney Dis 37:909-914, 2001 16. Hayslett JP, Krassner LS, Bensch KG, Kashgarian M,
ADULTHOOD OUTCOME OF CHILDHOOD NEPHROSIS
557
Epstein FH: Progression of “lipoid nephrosis” to renal insufficiency. N Engl J Med 281:181-187, 1969 17. Meyrier A, Noe¨ l LH, Auriche P, Callard P. Collaborative Group of the Socie´ te´ de Ne´ phrologie: Long-term renal tolerance of cyclosporin A treatment in adult idiopathic nephrotic syndrome. Kidney Int 45:1446-1456, 1994 18. Foote KD, Brocklebank JT, Meadow SR: Height attainment in children with steroid-responsive nephrotic syndrome. Lancet 2:917-919, 1985
19. American College of Rheumatology Ad Hoc Committee on Glucocorticoid-Induced Osteoporosis: Recommendations for the prevention and treatment of glucocorticoidinduced osteoporosis: 2001 Update. Arthritis Rheum 44: 1496-1503, 2001 20. McClung MR, Geusens P, Miller PD, et al: Hip Intervention Program Study Group. Effect of risedronate on the risk of hip fracture in elderly women. N Engl J Med 344:333-340, 2001