Steroid-Free Immunosuppression Since 1999: 129 Pediatric Renal Transplants with Sustained Graft and Patient Benefits


* Corresponding authors: Minnie Sarwal and Oscar Salvatierra,,


Despite early promising patient and graft outcomes with steroid-free (SF) immunosuppression in pediatric kidney transplant recipients, data on long-term safety and efficacy results are lacking. We present our single-center experience with 129 consecutive pediatric kidney transplant recipients on SF immunosuppression, with a mean follow-up of 5 years. Outcomes are compared against a matched cohort of 57 concurrent recipients treated with steroid-based (SB) immunosuppression. In the SF group, 87% of kidney recipients with functioning grafts remain corticosteroid-free. Actual intent-to-treat SF (ITT-SF) and still-on-protocol SF patient survivals are 96% and 96%, respectively, actual graft survivals for both groups are 93% and 96%, respectively and actual death-censored graft survivals for both groups are 97% and 99%, respectively. Unprecedented catch-up growth is observed in SF recipients below 12 years of age. Continued low rates of acute rejection, posttransplant diabetes mellitus (PTDM), hypertension and hyperlipidemia are seen in SF patients, with sustained benefits for graft function. In conclusion, extended enrollment and longer experience with SF immunosuppression for renal transplantation in low-risk children confirms protocol safety, continued benefits for growth and graft function, low acute rejection rates and reduced cardiovascular morbidity.


acute rejection


body mass index




Epstein–Barr virus




end-stage renal disease




granulocyte colony stimulating factor


intent-to-treat steroid-free


mycophenolate mofetil


North American Renal Transplant Cooperative Study


standard deviation scores


Chronic corticosteroid usage, a cornerstone drug for the last five decades in organ transplantation (1), is a major cause of morbidity, ranging from cardiovascular disease, hyperlipidemia, hypertension, cataracts, osteoporosis, weight gain, acne, diabetes mellitus, to bacterial and viral infections. Many of the steroid-related cosmetic side effects drive immunosuppression nonadherence, a particular problem in adolescence (2). Additionally, the suppressive effect of chronic, even low-dose steroids, on growth velocity in children is a critical concern (3,4). Steroid withdrawal has been attempted in various protocols over the last decade, and though successful in selected patients (5), can result in unacceptably high rates of acute rejection and even graft loss (6–10). Steroid minimization does not always eliminate all steroid-related side effects (11–13), which has led to an interest in developing protocols to avoid maintenance steroids in renal transplantation, under the umbrella of more powerful immunosuppressive agents for induction and maintenance therapy (14–18).

Recent single-center, nonrandomized studies from both the pediatric and adult experience with steroid avoidance have demonstrated reduced morbidity from cosmetic and hemodynamic steroid side effects, reduced viral infection (19) as well as potential benefits for graft function and acute rejection (4,20–30), but these results need to be reevaluated after the completion of randomized controlled trials. These reported experiences in renal transplantation with steroid avoidance span 5 years in adult renal patients (23,27) and 2 years in pediatric renal patients (24). In this analysis, we present extended enrollment and follow-up of 129 consecutive pediatric and infant recipients treated at a single center since November of 1999. Important questions addressed in this study were: selected comparisons with a concurrent matched steroid-based (SB) cohort, long-term tolerability of the steroid avoidance protocol, early and late acute rejection rates, cardiovascular risk factors, growth benefits, graft function and patient and graft survival.

Materials and Methods


This study reports on a single-center experience for all 129 low immunologic risk (<20% peak panel reacting antibody or PRA) pediatric (0–21 years of age) recipients of a primary kidney allograft at Stanford University transplanted between November 1999 and November 2007, on the Stanford steroid-free (SF) immunosuppression protocol. This previously reported protocol (4) utilizes tacrolimus, mycophenolate mofetil (MMF) and extended daclizumab induction for 6 months for a total 10 mg/kg versus 5 mg/kg for the standard 8 weeks daclizumab induction used in the SB comparative group (Figure 1). Extended daclizumab induction in this protocol is governed by an Investigational New Drug (IND) (BB-IND-10127) held by M. S. until 2004 and since then, by the National Institute of Allergy and Infectious Diseases (NIAID) at the National Institutes of Health (NIH), when this SF protocol was subsequently tested in an NIH-funded randomized multicenter study (SNSO1). Patients have a mean follow-up of 5 years (range 0.8–8.8 years). Patient numbers with complete follow-up at 1, 3, 5, 7 and 8 years are, respectively, 125, 96, 70, 48 and 29. The first 20 SF patients were maintained on higher tacrolimus target tough levels in the first posttransplant week (17–20 ng/mL) and higher dosages of MMF (600 mg/m2/dose, twice daily; without drug levels). With early protocol safety, tacrolimus targets were lowered (12–14 ng/mL for the first week posttransplantation) and MMF dose has been reduced by 50% to 300 mg/m2/dose twice daily (24,31). In case of MMF intolerance (gastrointestinal or bone marrow suppression) within the first 6 months posttransplant, MMF was temporarily replaced by azathioprine (Imuran®, GlaxoSmithKline, Brentford, Middlesex, UK) and patients were rechallenged with MMF prior to 6 months. Continued MMF intolerance after 6 months resulted in MMF being replaced by sirolimus (Rapamycin®, Wyeth, Madison, NJ). Sirolimus maintenance dose was 2–8 mg/day depended on patient weight and achievement of trough level of 5–8 ng/dL for the first year posttransplantation and 3–5 ng/dL the year after. Overall, 15% patients are on tacrolimus and sirolimus maintenance therapy and selected data subanalyses have been performed for this patient group.

Figure 1.

Stanford steroid-free immunosuppression protocol. Steroid-free immunosuppression usage is marked with the solid black line; steroid-based immunosuppression is marked with the hashed grey line. Protocol biopsies (denoted by stars) were performed at baseline (prior to reperfusion of the kidney), 3, 6, 12 and 24 months posttransplantation. Extended daclizumab induction was given for 6 months posttransplantation in SF patients (black solid line) versus standard usage of 2 months daclizumab induction in SB patients (gray dashed line).

SF recipients were compared with a matched control group (for recipient gender, age, race, and cause of end-stage renal disease [ESRD], height, weight, donor source and donor age) of 57 unsensitized recipients of primary transplants, treated with SB immunosuppressive protocol utilizing daclizumab induction for 2 months, tacrolimus and MMF (Table 1). SB patients were those who either did not consent to SF immunosuppression, or those who came to transplantation on low-dose steroids for their primary disease. The incidence of glomerulonephritis (41/129 or 32% in SF vs. 20/57 or 35% in SB) and focal segmental glomerulosclerosis (FSGS; 3/57 or 5.3% in SB vs. 7/129 or 5.4% in SF) are equivalent in the both treatment groups. In order to maximize the quality of the control group, the following selection criteria were also included: (1) As the SF protocol started in November of 1999, all SB patients in the control group were selected, on the basis of having received their transplant after this date, concomitant with the start of the SF protocol, making this concurrently transplanted group of patients; (2) peak PRA <20%, similar to the SF group; (3) primary kidney transplant recipients, similar to the SF group; (4) 100% 1-year graft survival; and (5) homogeneity of immunosuppression usage with all patients on tacrolimus, MMF and daclizumab induction, with the exception that the latter induction utilized the standard 2-month dosing regimen. Within the SB group 57 patients were enrolled with a mean group follow-up of 67 months. Prednisone 10 mg/kg was given perioperatively followed by 2 mg/kg/day in subjects weighing <40 kg and 1.5 mg/kg/day in subjects weighing >40 kg. The steroid dosing was tapered as follows: by the end of weeks 1, 2, 4, 6, 12 and 16, dosages were 0.5, 0.4, 0.3, 0.2, 0.15 and 0.1 mg/kg/day, respectively. There were similar proportions of children ≤ 6 years of age in both groups as recipients of an adult-sized kidney (29% in SF and 19% in SB, p = 0.30).

Table 1.  Patient demographics of SF and SB patients in this study
Baseline patient characteristics and immunosuppression
  1. 1Race: 1 = Caucasian; 2 = Hispanic; 3 = Asian; 4 = African American; 5 = Other.

  2. 2ESRD = end-stage renal disease: ESRD categories: 1 = glomerulonephritis; 2 = polycystic kidney disease; 3 = dysplasia; 4 = reflux nephropathy; 5 = obstructive uropathy; 6 = FSGS; 7 = others;

  3. PRA = panel reacting antibody levels.

Before transplantation12957 
 Mean age (yrs)11.0 ± 6.712.5 ± 6.10.18
 Mean age (yrs) males10.7 ± 6.812.6 ± 6.80.23
 Mean age (yrs) females11.7 ± 6.512.4 ± 5.70.65
 Gender: females39%55%0.07
 Mean height (cm)130.5 ± 35.9136.2 ± 35.00.34
 Height Z-score−1.2 ± 1.5 −1.1 ± 1.50.72
 Mean weight (kg) 38.2 ± 23.2 39.5 ± 20.80.74
 Mean BMI (kg/m2)20.3 ± 4.719.6 ± 4.20.35
 Mean BSA (m2) 1.2 ± 0.5 1.2 ± 0.50.54
 Race (1,2,3,4,5)154%, 21%, 9%, 8%, 8%52%, 24%, 12%, 4%, 8%0.85
 Cause of ESRD (1,2,3,4,5,6,7)232%, 10%, 16%, 13%, 18%, 5%, 6%35%, 6%, 2%, 8%, 12%, 5%, 32%0.1
 Dialysis pretransplant70%80%0.13
 PRA < 20%100%100%1
 Mean hematocrit (%)33.7 ± 6.633.7 ± 7.00.99
 Mean white cell count(uL) 8.4 ± 3.9 8.0 ± 3.50.54
 Mean serum cholesterol (mg/dL)161.5 ± 48.6135.9 ± 43.60.01
 Mean serum triglyceride (mg/dL)170.6 ± 95.0182.1 ± 91.00.57
 Hypertension (%)52%61%0.37
 Mean donor age34.4 ± 11.233.9 ± 10.50.84
 Donor age range (yr)2.3–570.6–50 
 Donor type (LRD)79%68%0.12
 Mean HLA match LRD2.8 ± 1.42.8 ± 1.30.99
 Mean HLA match CAD1.6 ± 0.92.0 ± 1.70.87

All SF and SB patients had protocol biopsies at 3, 6, 12 and 24 months after transplantation, and for graft dysfunction. All rejection episodes were biopsy proven and graded by the Banff classification (32). Clinical acute cellular rejection was treated with three pulses of intravenous corticosteroid (10 mg/kg) with the immediate return to SF maintenance therapy. Vascular rejection was treated with thymoglobulin with steroid premedication (1 mg/kg/dose) and consideration of breaking protocol to SB regimen (24). Subclinical borderline Banff grade rejection was treated with immunosuppression intensification without steroid pulsing. Steroids were introduced as maintenance immunosuppression if a rejection episode recurred within 3 months.

Graft function assessment was based on calculated creatinine clearance by the Schwartz formula (33,34). Height was measured in centimeters and presented as Z-scores (mean height standard deviation scores; mean height SDS), with normative data used for growth from the Center for Disease Control (CDC). Weight was measured in kilograms and body mass index (BMI) was calculated. Hypertension was defined as present if there was requirement of antihypertensive agents for clinical blood pressure control. Hypertriglyceridemia was defined as fasting serum triglyceride levels >140 mg/dL, hypercholesterolemia as fasting serum cholesterol levels >190 mg/dL and diabetes mellitus as hyperglycemia (fasting plasma glucose >126 mg/dL, based on the definition used by the American Diabetes Association), requiring treatment with either insulin or hypoglycemia agents. Hospitalizations were captured, with cause. This study was approved by the Institutional Review Board at Stanford University. To determine the effects of SF on nonspecific chronic tubulointerstitial damage, all protocol biopsies performed at 1 and 2 years posttransplantation were reviewed blindly by the same single pathologist. The extent of interstitial fibrosis/tubular atrophy (IF/TA) was scored on a scale of 0–3 based on the Banff criteria (32,35).

Statistical analysis

T-test, chi-square test, Fisher exact and analysis of variance (ANOVA) were used for analysis of continuous or categorical types of data. Linear mixed-effects models were used for evaluation of repeated measures for clinical outcomes comparison such as by adjusting time point categorical variables. For the binary outcomes we used repeated measures generalized estimating equations (GEE) model by adjusting time point categorical variables. Graft and patient death-censored survival analysis were based on Kaplan–Meier survival analysis and chi-square analysis. We used an epidermal growth factor receptor (eGFR) based on the Schwartz equation (33,34) of <10, or return to dialysis as an estimate of graft loss. Cox proportional hazard models were used to analyze hazard ratios by adjusting patient demographical variables for AR. Multivariate analysis and Pearson/Spearman correlation were used for examining relationship for all parametric or nonparametric variables. Multiple regression and logistic regression models were used to test associations with confounding clinical parameters. The p-values ≤ 0.05 were considered statistically significant. Results are reported as mean ± standard deviation. All statistical analyses were performed using the SAS 9.1.3 (SAS Institute Inc., Cary, NC).


SF protocol breaks

At mean follow-up of 5 years after transplantation, 87% of 129 SF patients remain steroid-free. Protocol breaks to SB maintenance therapy occurred in a total of 17 patients, with early breaks in 6 patients between 0 and 6 months and late breaks in 11 patients, between 6 months and 36 months posttransplantation. Causes for early protocol break were refractory AR (n = 3) and recurrent glomerulonephritis (n = 3). Causes for late protocol break were refractory AR (n = 6), recurrent glomerulonephritis (n = 3), PTLD (n = 1) and de novo reactive arthritis (n = 1) (24); these cases have been discussed in detail in a separate manuscript from our group (Sutherland et al., under review).

Patient and graft survival

Actual patient survival at the current follow-up was equivalent: 96% in ITT-SF patients, 96% in patients maintained on the SF protocol and 98% in SB patients (p = 0.42 for ITT-SF vs. SB and p = 0.51 for still on SF vs. SB). Four patients in the SF group died (3 in 2005 and 1 in 2002) with functioning grafts (mean serum creatinine for the group at the time of death was 0.8 mg/dL). The causes of death were: bacterial sepsis in a patient with an indwelling intravenous line, pretransplant diagnosis of severe combined immunodeficiency syndrome, hemophagocytic lymphohistiocytosis and accidental narcotic drug overdose. One patient died with a functioning graft in the SB group due to recurrent Wilm's tumor. Multivariate analysis was conducted for patient and graft survival across multiple clinical factors (immunosuppression protocol, donor age, donor source, cause of ESRD, recipient age and recipient gender), using a logistic regression model. There was no significant correlation of any of these confounders with graft or patient survival. Actual graft survival at current follow-up is 93% for all ITT-SF patients, 96% in patients still maintained on the SF protocol and 91% in SB patients (no differences in graft survival comparisons for on-treatment SF, ITT-SF and SB groups; Figure 2A, and B). The incidence of death-censored graft survival was better in the still-on-protocol SF versus SB patients (99% vs. 93%; p = 0.026), but not different when ITT-SF death-censored graft survival was compared with SB survival (97% vs. 93%; p = 0.25). The Kaplan–Meier survival analysis of these patient groups shows a similar significance for still-on-protocol SF (Figure 2C; p = 0.02), but no difference for ITT-SF (p = 0.40). There has been only one retransplantation in the SF group and this was because of progressive graft injury from drug nephrotoxicity. The patient received a preemptive second transplant from a parent at 8 years following the initial transplant. Despite a nonsignificant difference in living donor percentage between SF and SB groups (79% vs. 68%) there was no negative impact on graft outcomes in the SB group, based on donor source, as only one graft loss in SB group occurred in a patient receiving a kidney from a deceased donor. Also, even though there was an equal incidence of FSGS as a cause of ESRD in both groups, there was no graft loss or patient death due to recurrence of FSGS in either group.

Figure 2.

Graft survival and acute rejection results for SF and SB patients. (A) Graft survival for SF and SB patients. Kaplan–Meier survival analysis curves are shown for patients currently maintained on SF (n = 112; black line) and concurrently transplanted SB (n = 57; gray line) patients until current follow-up. (B) Graft survival for ITT-SF and SB patients. Kaplan–Meier survival analysis curves are shown for patients with ITT-SF protocol (n = 129; black line) and concurrently transplanted SB (n = 57; gray line) patients until current follow-up. (C) Death-censored graft survival for SF and SB patients. Kaplan–Meier survival analysis curves are shown for patients currently maintained on SF (n = 112; black line) and concurrently transplanted SB (n = 57; gray line) patients until current follow-up. (D) Incidence of acute rejection in ITT-SF, maintained on SF, and SB patients. This figure shows a comparison of the percentage of clinical acute rejection, late clinical AR (occurring after 1-year posttransplantation), early clinical AR (occurring within 1-year posttransplantation) among ITT-SF, (black color bar), maintained on SF (darker gray color bar), and only the SB (dark gray color bar) patients. The p-values were listed either between ITT-SF versus SB or maintained on SF versus SB. (E) Probability of remaining free from AR in SF and SB patients. This Kaplan–Meier analysis curves are shown for probability of remaining free of clinical rejection, between SF (black color bar), and the SB (dark gray color bar) patients at current follow-up.

Acute rejection (AR)

Clinical acute rejection:  The incidence of clinical AR in the first posttransplant year did not have any significant difference in patients ITT-SF (12%) versus on SB treatment (12%) (p = 0.98) or maintained on SF (7%) versus on SB treatment with p = 0.33 (Figure 2D); however, the overall incidence of clinical AR at study follow-up (mean of 5 years) was low in an ITT-SF (16%), and also low in patients maintained on SF (11%) (Figure 2D). There was no difference in the mean time to first AR (14 ± 13 months in either maintained on SF or 12 ± 12 months on ITT-SF vs. 26 ± 29 months in SB; p = 0.2 or p = 0.12), rejection-free survival was excellent for the SF group at current follow-up (log-rank p = 0.06, Figure 2E). Multivariate analysis of the characteristics of patients (cause of ESRD, recipient age, donor age and donor source) using a Cox proportional hazard model showed that the incidence of clinical AR did not associate with any of these adjusting clinical parameters.

Late clinical AR was defined as an acute rejection episode that occurred after the first year posttransplantation. There was a low rate of late AR in SF patients at 4% with a rate of 12% for SB. When stratified by years’ posttransplantation, the incidence of late AR (≥2 years) in an ITT-SF was 3%, and 10% in SB (Figure 2D). The mean time for late AR in ITT-SF patient (28 ± 10 months) versus SB (47 ± 29 months) was similar with p = 0.17, and maintained on SF versus SB was similar (30 ± 8 months) with p = 0.3.

Subclinical acute rejection:  Ten percent of ITT-SF patients and 4% of SB patients had subclinical AR. There was no difference in the incidence of subclinical acute rejection in SF patients based on protocol biopsies alone. Simultaneous with the reduction of maintenance target tacrolimus trough levels (24,31), the incidence of subclinical AR episodes increased (0.75 sub-clinical rejections per year prior to 2002, and two subclinical rejections per year after 2003).

Graft function

Graft function, as measured by the Schwartz calculated creatinine clearance in mL/min/1.73 m2 (33,34), was significantly higher at all measured annual time points in the ITT-SF versus the SB group (F = 28.9, mixed model group effect p < 0.0001, Figure 3A), regardless of recipient age (F = 17.9, mixed model group effect p = 0.0001; Figure 3B). Using a multiple regression model by adjusting clinical AR episodes, donor age, antihypertensive agents, serum triglyceride levels and serum cholesterol levels, the graft function benefit in the SF cohort at 1-year posttransplantation strongly correlates with a lower number of antihypertensive agents (p = 0.04), lower serum triglyceride levels (p = 0.07) and fewer clinical AR episodes (p = 0.03) (model F = 4.79, p = 0.0002).

Figure 3.

Comparison of graft function between ITT-SF and SB protocols. (A) For overall age of patients, calculated creatinine clearance (mL/min/1.73 m2; y-axis) is plotted against months of follow-up posttransplantation (x-axis) for ITT-SF (black line) and SB (gray line) protocols. Each dot with the solid vertical lines (up/down) represents mean ± standard deviation. F = 28.9, mixed model group effect p < 0.0001 was derived from linear mixed model by repeated measures. (B) For children (age ≤6 years), calculated creatinine clearance (mL/min/1.73 m2; (y-axis) is plotted against months of follow-up posttransplantation (x-axis) for ITT-SF (black line) and SB (gray line) protocols. Each dot with the solid vertical lines (up/down) represents mean ± standard deviation. F = 17.9, mixed model group effect p = 0.0001 was derived from linear mixed model by repeated measures.


SF recipients ≤6 years of age were enrolled in 2001, after safety data were evaluated for the first 10 older children (6–21 years; Reference 4), thus growth analysis is reported over 6 years so as to include all age groups. Catch-up linear growth trends, represented by Z-scores (mean height SDS; Figure 4), were highly significant in the youngest recipients in still-on protocol SF (n = 112) versus SB immunosuppression (F = 2.8, mixed model p = 0.01) (Figure 4A). After 4 years posttransplantation, the youngest SF children show unprecedented catch-up growth rates, greater than the growth velocity in normal healthy age and gender-matched controls (from the CDC). Height benefits were also seen for SF children 6–12 years of age, over the first 6 years posttransplantation (F = 2.9, mixed model p = 0.01; Figure 4B). Children ≥12 years of age at the time of transplantation, on SF treatment or ITT-SF, demonstrated improved growth trends, which do not reach significance in this study. When comparative growth analysis was performed for the ITT-SF (n = 129) versus the SB group, all significant differences in growth data persist for the SF group.

Figure 4.

Growth comparisons between SF and SB protocols. (A) Mean height Z-score were shown for children (age ≤6 years) for SF (black bar) and SB (gray bar) protocols. The values are shown under the figure. p = 0.01 was derived from the linear mixed model by repeated measures. (B) Mean delta height Z-scores are shown for young patients (age 6–12 years) for SF (black bar) and SB (gray bar) protocols. p = 0.01 was derived from the linear mixed model by repeated measures.

Cardiovascular risk factors

Diabetes mellitus:  The incidence of posttransplant diabetes (PTDM) was lower in the SF group: 0.8% versus 9%; p = 0.0089. Two African-American patients, who were originally on SF immunosuppression, developed PTDM only after breaking to SB maintenance therapy for recurrent rejection in one and recurrence of immunologic disease (Reiter's arthritis) in the other.

Hypertension and hyperlipidemia:  SF patients, maintained on the protocol, had significantly lower requirement for antihypertensive agents at all time points posttransplantation (chi-square = 19.86, GEE model group effect p < 0.0001); with a significantly higher prevalence in SB patients (Table 2). Serum triglyceride levels were also statistically lower in on-protocol SF patients (F = 8.7, mixed model group effect p = 0.0038; Table 2). Irrespective of sirolimus maintenance therapy in 17 SF patients, and the fact that at baseline, SF patients happened to have statistically higher serum cholesterol levels (Table 1), significantly lower serum cholesterol levels were seen in SF patients compared to SB patients at all time points measured (F = 4.9, p = 0.0285) (Table 2). There was a similar incidence of hyperlipidemia in SF patients on tacrolimus and rapamycin (45%) at 1-year posttransplantation when compared to SB patients who were on tacrolimus, MMF and steroids (40%). In the SF group, hyperlipidemia was mild, even with sirolimus therapy and required treatment with lipid-lowering agents in only 2% of SF patients (all on sirolimus), whereas 80% of the SB patients had hyperlipidemia (in the absence of sirolimus maintenance therapy).

Table 2.  Posttransplantation hypertension and hyperlipidemia
Posttransplantation hypertension and hyperlipidemia
  1. 1Hypertension was defined as present if there was requirement of antihypertensive agents for clinical blood pressure control.

  2. 2Hypercholesterolemia was defined as fasting serum cholesterol levels >190 mg/dL.

  3. 3Hypertriglyceridemia was defined as fasting serum triglyceride levels >140 mg/dL.

Hypertension (%)1
 1 year31%55%<0.0001
 3 years12%70% 
 5 years14%50% 
 7 years8%56% 
Hypercholesterolemia (%)2
 1 year15%40%0.0285
 3 years18%26% 
 5 years20%50% 
 7 years25%75% 
Hypertriglyceridemia (%)3
 1 year45%60%0.0038
 3 years32%65% 
 5 years26%50% 
 7 years8%80% 

Body mass index:  SF patients showed smaller changes in BMI after transplantation, compared to SB, but significant changes occurred only in the first 2 years posttransplantation (Δ BMI at 1 year 1.5 ± 3.5 in SF vs. 3.3 ± 3.7 kg/m2 in SB; p = 0.0034, and Δ BMI at 2 years 1.6±6.1 in SF vs. 3.8 ± 3.9 kg/m2 in SB; p = 0.02). Approximately 20% of patients were overweight (BMI >25 kg/m2) in SF arm and 30% in SB arm at 1 and 2 years of follow-up, and approximately 10% were obese (BMI > 30 kg/m2) patients in both arms at 1 and 2 years of follow-up.

Bone marrow suppression

In the first year post-transplantation, 18% of SF patients required usage of granulocyte colony-stimulating factor (GCSF) to support their white blood cell counts (WBC) within the normal range (4.0–11.0 × 109/L), while SB patients maintained their WBC counts without GCSF requirement (p = 0.0025). Most of the SF patients (82%) who required GCSF were transplanted earlier in the study, prior to 2002, when higher doses of MMF were used in the protocol (24). When stratifying the effect of higher MMF usage on bone marrow suppression in the first 20 patients (receiving 1200 mg/m2/day MMF) versus the last 20 patients enrolled (receiving 600 mg/m2/day MMF) with at least 1-year follow-up, there was less GCSF usage (60% vs. 8%; p = 0.004), with similar WBC counts (7.4 ± 2.7 in the first 20 vs. 7.4 ± 1.9 × 109/L in the last 20; p = 0.98) when lower MMF doses were used. The hematocrit was higher in SF group in the first posttransplant year (37%± 5% vs. 34%± 4%; p < 0.0001), without more erythropoetin (EPO) usage in SF patients early posttransplantation (25% vs. 18%; p = 0.34). After 2 years posttransplant, there were no significant differences in HCT, WBC counts, EPO and GCSF usage.


There was no increased incidence of bacterial infections or hospitalizations in the SF group (32% in SB vs. 36% in SF, p = 0.6). Subclinical EBV viremia, measured by peripheral blood quantitative PCR, was detected in 22% of SF patients and 57% of SB patients (p < 0.05), but there were no differences in the incidence of subclinical CMV viremia (12% in SF vs. 17% in SB; p = 0.1). We have previously reported on the time to detection of subclinical EBV and CMV viremia and its clinical impact in SF and SB patients (19). There was a low incidence of CMV disease (2 SF patients vs. 1 SB patient), 1 nodal PTLD in SF and 4 nodal PTLD in SB (p = 0.17), and 2 SF patients were diagnosed with BK nephritis.

Protocol biopsy results and incidence of chronic graft injury

Within the SF group, 76 protocol biopsies at 1 year and 47 biopsies at 2 years were reviewed and compared with the SB group with 34 biopsies at 1 year and 17 biopsies at 2 years. The IF/TA score in all the groups ranged from 0 to 3 and there was no significant difference in the IF/TA score between the two immunosuppression groups either at 1 (p = 0.1) or 2 years posttransplantation (p = 0.46).


We have herein reported on an extended experience with SF immunosuppression in pediatric renal transplantation. This single-center study provides the largest experience and the longest outcome data following transplantation without any planned steroids in children. Eight-seven percent of patients were able to be maintained on SF immunosuppression at a mean of 5 years (0.8–8.8 years) of study follow-up. The most important observation of this study is that steroid avoidance with maintenance tacrolimus and MMF and extended induction with daclizumab for 6 months is safe, without increased risk of infections and malignancies and is effective, with an increased tendency to stay rejection-free, slower decline in graft function and potentially improved long-term graft survival. Additional long-term benefits of dramatic catch-up growth trends are seen in young children, supporting the transplantation of children with renal failure before 12 years of age, accompanied by simultaneous avoidance of maintenance steroids. The important benefits of steroid avoidance on mitigating diabetes and cardiovascular risk factors are also appreciated.

An unexpected initial finding of this study, which persists on longer follow-up, is that the occurrence of acute clinical rejections was low in the milieu of complete steroid avoidance, with no planned pre-, intra- or post-operative steroid exposure. In the NAPRTCS 2007 report data (36), where data were collected on pediatric renal transplant outcomes between 2001 to 2006, the clinical acute rejection rate is 20.7% for deceased donor recipients and 24.7% for living donor recipients. The low immunological risk with this protocol may relate to many reasons. Low acute rejection rates have also been reported in the adult renal transplant experience from Matas et al. (27), where steroids were rapidly discontinued by the postoperative day 6 and in the CARMEN study (37) where there was complete steroid avoidance. Some studies employing steroid discontinuation many months after transplantation have reported a higher incidence of acute rejection (7,8). It has been hypothesized that as steroids suppress the release of many cytokines (38,39) and may inhibit peripheral tolerance via a Fas-dependant mechanism (40), continuation of steroids months after transplantation may result in ‘steroid dependency’, with an increased incidence of acute rejection on late steroid withdrawal. The FREEDOM (41) study, on the other hand, utilizing early steroid discontinuation or avoidance in adult renal transplantation, reports higher rejection rates with two doses of Basiliximab induction and cyclosporine-based maintenance immunosuppression, suggesting that effective induction and maintenance immunosuppression are both important for successful steroid avoidance. In addition, a recent randomized double-blind trial of early steroid cessation also shows the biopsy-confirmed AR approached significance for steroid withdrawal compared with chronic steroid therapy (42). Second, our study is restricted to low-risk, unsensitized recipients (PRA <20%), with a vast majority (69%) receiving living donor organs, while the deceased donors generally accepted at Stanford have short cold ischemia times (<18 h). Additionally, only 6% of SF recipients were African Americans. We wish to emphasize that comparisons regarding acute rejection rates were conducted with a SB group of similarly low immunologic profile. Third, the low acute rejection rate in this study may be due to the immunosuppression combination employed by the protocol: dual-maintenance immunosuppression, rather than monotherapy (43), which is tacrolimus based (27,44,45) and employs a novel prolonged IL-2 receptor blockade. The extended daclizumab usage in this protocol is likely a critical anchor in providing good immunologic protection at antigen presentation and the ensuring early months following transplantation. We and others have previously shown that the double pretransplant dose of 2 mg/kg provides serum daclizumab levels of >5 mcg/mL in all recipients ages (4,24,46), which may be crucial for suppressing IL2-dependant proliferative responses, independent of the saturation of the IL2-receptor (47). Finally, SF patients have a lower rate of CMV and EBV subclinical viremia (19), which could also reduce the immunologic risk of acute rejection.

The Canadian multicenter transplant study groups have expressed some concern for late allograft loss (22) with steroid avoidance, but the immunosuppression protocols in the study group vary considerably from the one used in this study. In an analysis of the subclinical histological evolution of the protocol biopsies, without any subclinical or clinical acute rejection, performed within the SF group, we reported on a significant increase in chronic tubulointerstitial injury in the first 2 years after transplantation unrelated to rejection (48). Donor organ–recipient size discrepancy appeared to be the most important determinant of this chronic tubulointerstitial damage, even in the absence of acute rejection. Comparisons of chronic graft injury scores between the SF and the SB group in this study showed a progressive increase in chronicity, but there were no differences in the incidence or extent of chronic graft injury in the two protocol groups. As the pathologist was blinded with regards to patient steroid usage, this analysis suggests that, rather than use or avoidance of steroids, other factors such as maintenance therapy with a calcineurin inhibitor, may be driving the progression of chronic graft injury after transplantation. The low rates of late acute rejection argue against increased long-term graft damage by the omission of steroids from the immunosuppressive protocol. This hypothesis appears to be validated in the comparative benefit for graft function and projected long-term graft survival for the SF over the SB groups. In addition, the mitigation of cardiovascular risk factors, such as hypertension, hyperlipidemia and diabetes in the SF group, will likely result in long-term patient benefits and also contribute to some observed long-term graft survival benefits. Notable among these, is the continued excellent overall graft function during the extended period of follow-up, showing that the steroid avoidance protocol is safe after discontinuation of immunologic support from the extended daclizumab induction. Importantly, the hyperlipidemia in the SF recipients was mild, even when the protocol utilized sirolimus maintenance therapy, whereas concomitant administration of sirolimus and steroids often resulted in significant hyperlipidemia in SB patients, even with low-dose maintenance steroids of 0.05–0.1 mg/kg/day. Similarly, the diabeteogenic effect of tacrolimus and other patient risk factors such as race may be greatly accentuated by concomitant steroid exposure. Thus, the patient tolerability of both tacrolimus and sirolimus may be improved in steroid-avoidance protocols.

An important patient benefit from this study is the demonstration of significant catch-up growth potential in children 1–12 years of age, after successful transplantation, previously masked by chronic steroid usage. We now observe that children from 1 to 6 years of age at the time of transplantation, in particular, start to show catch-up growth rates, even more than their age- and gender-matched peers. This signals that with steroid avoidance regimes, young children may be able to realize their final adult growth potential, but this would be influenced by the residual effects of uremia in childhood prior to transplantation. This catch-up growth benefit has not been previously reported in any other pediatric transplant study on steroid minimization. The absence of steroids is also important to realize the postpubertal growth potential in children transplanted at 6–12 years of age. However, children with renal transplantation over the age of 12 years, with sometimes important and long-standing growth failure from longer duration of chronic renal insufficiency prior to transplantation, do not show the same significant growth benefits as children from 1 to 12 years of age. Thus, preemptive transplantation in young children and infants prior to the age of 12, and avoiding steroids after transplantation, may provide the best window of opportunity to achieve full growth potential in children.

The early SF immunosuppressive protocol experience induced more bone marrow suppression than the SB protocol, which was likely related to the fact that in this phase of the protocol, higher MMF doses were used. In addition, the use of corticosteroids leads to leucocyte demargination, which could mask the effect of MMF use on the bone marrow. MMF doses are better tolerated at lower doses in SF patients, and can likely be safely lowered, as seen in this study, perhaps because of previously reported increased bioavailability of MMF in the absence of steroids (49).

A limitation of this study is that it is a single-center, nonrandomized analysis on a selected group of patients with low sensitization risk. Recognizing this and to assure as best as possible a robust control group, all control patients were required to have a functioning graft at 1-year posttransplantation. Our study benefits from being protocolized by a small and consistent physician group. As the early results of the study were encouraging (4,24), the protocol was rapidly accepted as standard of care at our center and randomization could not be done. However, the protocol was adopted as part of a multicenter US and Canadian NIH-funded study (SNSO1), including 130 patients randomized to the Stanford SF protocol with extended daclizumab induction versus ‘classic’ SB immunosuppression with standard 2 months daclizumab induction. Both arms were maintained on long-term tacrolimus and MMF therapy. We await the future report of the results of this study, which will also provide direct comparison of clinical results and chronic graft injury scores from protocol biopsies in both the SF and SB cohorts.

The adult renal transplant experience has also reported successful prednisone-free maintenance immunosuppression in higher-risk groups—for example, African Americans, those with high PRA levels and those with DGF (50,51), suggesting that steroid elimination may possibly be applicable to most pediatric recipient groups. However, additional study with larger numbers of African-American and deceased donor recipients, as well as repeat transplants, are necessary to test extension of this protocol to all pediatric recipient groups. The SNSO1 study will also be able to address this issue as it incorporates a larger number of African-American and deceased donors.

In summary, this study demonstrates the safety of using an SF immunosuppressive regimen for kidney transplantation in low-risk pediatric recipients. This protocol demonstrates that the incidence of acute rejection is no higher than that seen in SB protocols. Steroid avoidance appears to result in a growth advantage in recipients less than 12 years of age at the time of transplantation. In addition, there appears to be a mitigation of hypertension, hyperlipidemia, posttransplant diabetes and subclinical CMV and EBV viremia, which may be critical in driving excellent graft function observed in the SF patients. The apparent diminution of steroid-related side effects, in combination with a low acute rejection rate and good graft survival, suggest that transplant programs could consider eliminating steroids from immunosuppressive protocols for low-risk pediatric recipients.


We are grateful for the help from our pediatric transplant nephrology team, especially Dr. Scott Sutherland, Dr. Lauren Weintraub, the transplant nephrology fellows and for the invaluable and persisting efforts of all people involved in the pediatric kidney transplant program at Stanford University, namely, the clinicians, renal nurses, transplant coordinators, children and their families.

The funding source is Beta Sigma Phi Kidney Research Foundation and Lucile Packard Children's Hospital Foundation to M. S. and O. S.