Journal Logo

Editorials and Perspectives: Special Feature

Steroid Avoidance or Withdrawal After Renal Transplantation Increases the Risk of Acute Rejection but Decreases Cardiovascular Risk. A Meta-Analysis

Knight, Simon R.1,2; Morris, Peter J.1,3

Author Information
doi: 10.1097/TP.0b013e3181c518cc
  • Free

Abstract

Corticosteroids have been an important part of maintenance immunosuppression protocols since the early 1960s, when they were initially used to supplement the immunosuppressive properties of azathioprine (AZA) (1). Although they are undoubtedly effective at preventing rejection in transplant recipients, the side effect profile of steroids has resulted in attempts first to reduce the dose of steroids and second with the advent of newer more potent immunosuppression to avoid or withdraw steroids to improve long-term safety outcomes.

Of particular interest are the effects of steroids on lipid and glucose metabolism and blood pressure regulation. Death with a functioning graft is an important cause of graft loss after renal transplantation, accounting for more than 40% of graft losses in one registry analysis (2). The leading cause of death with a functioning graft in this analysis of renal transplant recipients was cardiovascular disease, and therefore, there is a strong case for steroid minimization regimens that attempt to reduce long-term cardiovascular risk. Furthermore, hyperlipidemia, hypertension, and diabetes have all been identified as factors directly implicated in the pathogenesis of chronic allograft nephropathy (3, 4). However, any advantage seen in cardiovascular risk must be balanced against the potential reduction in immunosuppression and the risk of acute rejection (AR), which in turn is linked to graft loss and dysfunction that would make withdrawal or avoidance of steroids less attractive.

Another aspect of the many complications of steroid therapy is the potential increase in healthcare costs that they incur. An analysis by Veenstra et al. in the United States has demonstrated that the 10-year cost of treating steroid-related side-effects after transplantation is an additional US $5300 per patient. The bulk of this additional cost is in treating hypertension ($1878), closely followed by posttransplant diabetes mellitus ($1794) (5).

A number of trials have investigated the avoidance or withdrawal of steroids alongside various maintenance immunosuppression regimens over a number of years, with variable results. One of the main problems with such trials is that they rarely have the statistical power to identify differences in uncommon outcomes, such as graft loss and patient death, between the arms. A number of authors have attempted to address this issue by performing meta-analyses of trials of steroid avoidance or withdrawal (SAW) in renal transplant recipients (6–11). These analyses suggest that the risk of AR is increased significantly with steroid withdrawal or avoidance, although the impact on long-term graft function and survival is less clear with only one study demonstrating increased graft failure with steroid withdrawal (7). Only one study with restrictive inclusion criteria investigated cardiovascular risk factors after steroid withdrawal, demonstrating significantly lower serum cholesterol in the withdrawal group (8, 9).

Further conflicting evidence has been provided from nonrandomized registry data. In a retrospective analysis of data from the Collaborative Transplant Study (CTS), Opelz (12) suggested that graft survival was significantly improved during 5-years in renal transplant recipients in whom steroids were withdrawn from protocols including cyclosporine with or without AZA. This prompted a prospective study in which renal and cardiac transplant recipients were withdrawn from steroids beyond 6-months after transplantation and compared with matched controls from the CTS registry (13). This study again demonstrated an improvement in graft survival in the steroid withdrawal patients at 7 years (81.9% vs. 75.3%, P=0.0001), along with an improvement in patient survival at the same time point (88.8% vs. 84.3%, P=0.0016). In contrast to the meta-analyses described earlier, rates of AR and graft dysfunction did not differ between groups. Some evidence was found for an improvement in cardiovascular risk factors after steroid withdrawal.

Clearly, there is still considerable controversy regarding the risks and benefits of withdrawing steroids from maintenance steroid regimens. A number of more recent studies have provided further information and some of the studies included in previous analyses have reported long-term follow-up data. This study will comprise a comprehensive systematic review and meta-analysis of steroid avoidance and withdrawal protocols across all maintenance immunosuppression regimens in renal transplant recipients, with a particular emphasis on investigating the impact on long-term graft function and survival and with the impact on cardiovascular risk factors.

METHODS

Literature Search and Inclusion Criteria

A systematic literature search was performed using OVID MEDLINE and EMBASE, the Cochrane Central Registry of Controlled Trials, the Transplant Library from the Centre for Evidence in Transplantation (which includes hand-searched journals and conference proceedings), and trial registries (clinicaltrials.gov, the national research register, and current controlled trials). To avoid missing potentially relevant references, searches were performed using only Mesh keywords and free-text aliases for corticosteroids in each database, without limiting searches further using terms for sparing and avoidance. No date or language limits were applied. References of included studies and previous relevant reviews were scanned for potentially relevant studies that had been missed in the literature searching. The final date for literature searches was 18th January 2008. A limited updated search of the Transplant Library was performed from January 2008 to 18th August 2009 to identify any further studies published during preparation of this article.

Inclusion criteria specified any prospective randomized or pseudorandomized study in adult renal transplant recipients, in which outcomes in patients receiving maintenance steroids from the time of transplantation were compared with a cohort in which steroids were withdrawn at any time posttransplant or were avoided. Those studies investigating nonrenal transplants and pediatric recipients are considered elsewhere (14). Studies in which steroids were used for other conditions were excluded. Studies in which steroid doses were minimized, but not withdrawn, were also excluded.

Outcomes

The primary outcome in this analysis was the incidence of AR. Biopsy-proven AR was used for analysis where reported, otherwise the study author's definition of AR was accepted. Secondary outcomes were patient and graft survival, hypertension, diabetes, hypercholesterolemia, infection, malignancy, cataracts, and bone complications.

Data Abstraction

Studies are referred to throughout this article by the first author and year of the first peer-reviewed publication from that study. In the absence of any peer-reviewed publications, the first author and year of the first published abstract is used. Demographic, quality, and outcome data were extracted from the included studies into a custom-designed online database by the lead author (S.R.K.). The quality of the extracted data was confirmed by double checking by the second author (P.J.M.). Disagreements were resolved by discussion. There was also continuous cross-checking of previous entries during the data analysis.

Quality Assessment

Quality was assessed both by means of the Jadad score (a score of 3 or greater is considered good quality) and a description of allocation concealment and analysis based on intention-to-treat (15). When assessing study quality, all reports from a trial are assessed and the information pooled.

Data Synthesis

Summary effects were calculated in each study for outcomes using the meta and rmeta packages for the R statistical analysis language (16). For binary outcomes, the relative risk (RR) was used as a summary statistic, with the weighted mean difference (WMD) used for continuous outcomes. For survival outcomes, the hazard ratio (HR) was used. All summary effects are presented with a 95% confidence interval. If an outcome is reported at more than one time point for a single study, the most recent follow-up data are used.

Meta-analysis was performed using the same statistical software. Heterogeneity was quantified using the I2 test, which describes the percentage of total variation across the studies that is due to heterogeneity rather than chance (17). In the absence of heterogeneity, studies were combined using a Mantel–Haenszel fixed effects meta-analysis. If visual inspection of the forest plot or a high I2 value suggested heterogeneity, potential causes were explored using subgroup analyses, mixed-effects models, funnel plots (to search for evidence of publication bias), and by looking for methodological differences between the studies. If no explanation could be found, a DerSimonian and Laird random effects meta-analysis was performed.

One study was identified with three arms—steroid avoidance, steroid withdrawal, and steroid maintenance (18–24). In subgroup analyses by withdrawal time, the control group (steroid maintenance) was split evenly for comparison with the withdrawal and avoidance groups to prevent duplication of patients in the overall analysis.

For the analysis of patient and graft survival data, the HR was used as it takes into account those patients who were lost to follow-up as well as the time to patient or graft loss. Unfortunately, few articles in transplantation report HRs for survival outcomes. Therefore, the methods of Parmar et al. (25) were used to estimate the HR from the reported data.

Missing Data

For a number of trials, data regarding continuous outcomes were poorly reported. Unless there was evidence of a non-normal distribution of results (skewed range), the median was taken as an estimate of the mean where the latter was not reported. The most common problem was the absence of a standard deviation or standard error. If a range and number of patients was reported, the methods of Walter and Yao (26) were used to estimate the standard deviation. Otherwise, a pooled standard deviation was calculated for all similar studies in the current meta-analysis and applied to studies with missing values (27). A sensitivity analysis was performed with and without studies requiring estimated standard deviations to assess their impact.

Subgroup Analyses

The time of steroid withdrawal after transplantation was regarded as likely to be an important factor in the outcome of steroid withdrawal protocols, and so a number of subgroups were defined accordingly:

  • Avoidance: No steroid induction or maintenance steroids given (but allowable for treatment of rejection).
  • Induction: 7 days or less of steroid therapy (includes single intraoperative steroid bolus).
  • Early withdrawal: Withdrawal of steroids from 8 days to 12 months after transplantation.
  • Late withdrawal: Withdrawal of steroids at any time from 12 months posttransplantation.

Another possible factor influencing outcome was the concomitant maintenance immunosuppression and the use of antibody induction therapy, and thus subgroup analyses were planned accordingly. Subgroups were compared using the interaction method described by Altman and Bland (28), with a significance level of P less than 0.05.

To further investigate the role of these potentially interacting factors, in particular when residual unexplained heterogeneity remained, attempts to fit a mixed-effects model were made using the MiMa function for the R language (29). This model allows linear (e.g., withdrawal time) and binary (e.g., presence of intention-to-treat analysis, use of antibody induction) moderator variables to be incorporated into a random-effects model. This then allows the amount of residual heterogeneity remaining (after accounting for the amount expected from the moderator variables) to be calculated (the QE statistic), along with the probability that at least one of the moderator variables significantly influences the effect sizes (the QME statistic). The P values for the individual moderator variables can also be calculated. This model has the advantages that linear variables such as withdrawal time can be included in their original form, rather than arbitrarily dividing studies into subgroups, and allows the effects of more than one variable to be analyzed in the same model.

RESULTS

Initial literature searches identified 10,997 references across all databases (OVID MEDLINE, 2614; OVID EMBASE, 5569; Cochrane Central Registry, 1751; and Transplant library, 1063). After removal of duplicates, this number was reduced to 7073 articles (Fig. 1). After exclusion on the basis of titles and abstracts, 497 papers were retrieved for full-text review. A further 13 potentially relevant trials were identified from the trial registries, and authors contacted for further information regarding the fate of the trials. Authors of studies only reported as abstracts with no full publication were also contacted for further information. The limited extended search to August 2009 identified a further two relevant publications—both articles providing long-term follow-up of previously identified studies (30, 31).

FIGURE 1.
FIGURE 1.:
Flow of studies through the review process. RCT, randomized controlled trial.

One hundred nineteen publications from 34 studies met the inclusion criteria in full. These included a total of 5637 patients. Details of the studies, including concomitant immunosuppression, time of steroid withdrawal, and quality data are shown in Table 1.

TABLE 1
TABLE 1:
Demographics and quality assessment of studies investigating steroid avoidance or withdrawal in adult renal transplant recipients

In the study by Kumar et al. (32), randomization was ceased early and a steroid withdrawal protocol adopted in all patients after review of interim results. Only reports from the randomized patients before study cessation are included in this review.

The methodological quality of the randomized controlled trials in this meta-analysis ranged from poor to good. Only 15 of the 34 studies (44%) achieved a Jadad score of 3 or greater, 15 studies (44%) reported an intention-to-treat analysis and 13 (38%) reported an adequate method of allocation concealment (Table 1).

Acute Rejection

Thirty-one studies (4626 patients) reported the incidence of AR, although two reported no events in either group and therefore had no weight in analysis (120, 123). Definitions of rejection varied, the majority (19 of 29, 66%) reported incidence of biopsy-proven AR, with the remainder reporting clinically defined AR (3 of 29, 10%) or giving no definition (7 of 20, 24%). Overall, the steroid avoidance and withdrawal regimens increased the risk of AR over controls (random effects analysis, RR 1.56, 95% confidence interval [CI] 1.31–1.87, P<0.0001; Fig. 2). Heterogeneity was high (I2=54.7%). Tests for interaction between subgroups and mixed-effects analysis incorporating withdrawal time as a linear moderator variable suggest that little of this heterogeneity results from different times of steroid withdrawal: the effect of withdrawal time on effect size was not significant (P=0.60), and residual heterogeneity was significant (P=0.0004). Subgroup analyses also demonstrate that the type of calcineurin inhibitor, antiproliferative agent, or the use or type of antibody induction do not explain the heterogeneity seen between the studies, nor significantly affect the RR of AR with steroid withdrawal (data not shown).

FIGURE 2.
FIGURE 2.:
Forest plot to show the RR of acute rejection with steroid avoidance or withdrawal at various times after transplantation. (blue boxes) Treatment effect for individual studies, (red diamonds) summary treatment effect for each subgroup and overall analysis derived by random effects analysis, and (horizontal lines) 95% confidence intervals. RR, relative risk; n, no. patients with acute rejection; N, total no. patients in study arm; CI, confidence interval.

In an attempt to explain the residual heterogeneity, an analysis was performed with just the high-quality studies (see Table 1). Random effects meta-analysis gives an RR of 1.68 (14 studies, 3102 patients, CI 1.30–2.17, P=0.001). Heterogeneity is actually increased, suggesting that study quality does not explain this finding (I2=65.7%). The funnel plot is grossly symmetrical, suggesting that there is no publication bias.

Corticosteroid-Resistant AR

Incidence of corticosteroid-resistant AR was reported in 13 studies (2797 patients). Of these, three studies reported no events in either arm (81, 94, 130) and therefore held no weight in meta-analysis. Overall risk of steroid-resistant AR did not differ significantly between the two groups (fixed effects, RR 0.95, CI 0.70–1.30, P=0.75, I2=0%). Asymmetry of the funnel plot demonstrates the possibility of some publication bias in favor of studies reporting a reduction in risk with avoidance/withdrawal.

Patient and Graft Survival

Twenty-nine studies reported patient survival data, but two (32, 60) reported insufficient data for estimation of the HR, and four reported no deaths in either group (92, 108, 123, 130). The overall HR for death was 0.82 (4650 patients, fixed effects, CI 0.61–1.11, P=0.2) with minimal heterogeneity (I2=0%). Time of withdrawal did not have an effect on patient survival after steroid withdrawal or avoidance. No difference in outcome was observed when only the high-quality studies were included in the analysis (13 studies, 3507 patients, HR 0.79, CI 0.56–1.11, P=0.18, I2=0%). Funnel plot demonstrated no obvious asymmetry.

Graft loss including death with a functioning graft was reported in 27 studies, although three studies reported insufficient data to enable HR estimation (60, 63, 130) and two reported no events in either group (92, 123). The overall HR for graft loss including death with a functioning graft is 0.99 (22 studies, 3790 patients, fixed effects, CI 0.80–1.22, P=0.93, I2=0%). Again, the funnel plot is grossly symmetrical.

Twenty-seven studies reported data regarding death-censored graft loss (excluding death with a functioning graft). Five of these studies reported no graft losses in either arm (73, 81, 92, 123, 142) and so had no weight in the meta-analysis. Two studies reported insufficient data to allow estimation of the HR (63, 130). Overall, no significant difference in death-censored graft loss was seen between maintenance and withdrawal or avoidance groups (20 studies, 4414 patients, fixed effects, HR 1.19, CI 0.92–1.54, P=0.19). No significant heterogeneity was observed (I2=0%). When only high-quality studies are included in the analysis, there is a trend toward an increased hazard of graft loss with SAW, but this remains nonsignificant (fixed effects, 10 studies, HR 1.31, CI 0.98–1.77, P=0.07, I2=0%). There is a suggestion of publication bias in the funnel plot, with an over-emphasis on studies reporting lower HRs. Therefore, the true hazard from withdrawal or avoidance may be higher than that estimated by the present meta-analysis.

No significant interaction is seen between time of withdrawal or concomitant immunosuppression and hazard for patient death or graft loss.

Graft Function

Serum creatinine was reported as an outcome in 26 studies. Where enough data were available, a pooled SD for all studies reporting withdrawal of steroids in the same time period was used for those studies not reporting a measure of variance. However, only one study in the steroid avoidance group (50) reported a standard deviation, and so a pooled estimate could not be calculated leading to exclusion of the other two studies in this subgroup from the meta-analysis (Spanish Monotherapy Study Group [52] and the steroid avoidance arm in the study by Vincenti et al.[24]). This left 25 studies (4101 patients) in the meta-analysis, which demonstrated a significantly higher serum creatinine in the patients undergoing steroid withdrawal or avoidance (fixed effects, WMD 4.24 μmol/L, CI 2.08–6.40, P=0.0001, I2=25.3%; Fig. 3). The funnel plot is grossly symmetrical.

FIGURE 3.
FIGURE 3.:
Forest plot to show the difference in serum creatinine with steroid avoidance or withdrawal. (blue boxes) Treatment effects for individual studies, (red diamond) summary treatment effect for overall analysis derived from fixed effects model, and (horizontal lines) 95% confidence intervals. Units are μmol/L. WMD, weighted mean difference, in μmol/L; SD, standard deviation; N, no. patients in study arm; CI, confidence interval.

Fourteen studies were identified that reported renal function as creatinine clearance. One further study reported function as glomerular filtration rate (24)—this was assumed to measure the same underlying effect and therefore the results were included in the meta-analysis. Fixed-effects analysis demonstrates a significant reduction in creatinine clearance across all studies with SAW (15 studies, 3016 patients, WMD −3.06 mL/min, CI −4.66 to −1.45, P=0.0002; Fig. 4). Heterogeneity is low (I2=13.6%). Funnel plot demonstrates some asymmetry, suggesting a degree of publication bias favoring studies with a smaller reduction or increase in creatinine clearance.

FIGURE 4.
FIGURE 4.:
Forest plot to show the difference in creatinine clearance with steroid avoidance or withdrawal. (blue boxes) Treatment effects for individual studies, (red diamond) summary treatment effect for overall analysis derived from fixed effects model, and (horizontal lines) 95% confidence intervals. Units are mL/min. WMD, weighted mean difference, in mL/min; SD, standard deviation; N, no. patients in study arm; CI, confidence interval.

No significant relationship between time of withdrawal or study quality and graft function was identified in subgroup analyses. Sensitivity analyses excluding those studies in which missing data were estimated (e.g., use of pooled standard deviations) are consistent with the overall data.

Cardiovascular Risk Factors

The effects of steroid avoidance and withdrawal on cardiovascular risk factors are summarized in Table 2. Overall, significant reductions in the risk of hypertension, new-onset diabetes, and hypercholesterolemia are seen with avoidance or withdrawal (Table 2). When only studies reporting an intention-to-treat analysis are included, the reductions in RR are smaller but remain significant (Table 2).

TABLE 2
TABLE 2:
Meta-analysis of cardiovascular risk factors in all studies and studies reporting intention-to-treat analysis only

Significant heterogeneity is found between studies in the analysis of hypercholesterolemia. A large proportion of the heterogeneity can be explained by the time of steroid withdrawal. Fitting of a mixed-effects model to the effect sizes incorporating the time of withdrawal as a linear moderator variable demonstrated no significant residual heterogeneity (test for residual heterogeneity QE=12.2, P=0.35) and identifies time of withdrawal as a significant moderator variable (z=−2.91, P=0.004). Therefore, the reduction in risk of hypercholesterolemia seems to increase with later steroid withdrawal.

No interaction is seen between concomitant immunosuppression or study quality and risk of hypertension, hypercholesterolemia, or new-onset diabetes. Funnel plots for all three outcomes demonstrated some asymmetry, suggesting publication bias in favor of studies demonstrating a larger reduction in risk with SAW.

Quantitative data regarding serum cholesterol levels were reported in 18 studies. Of these, four studies failed to report measures of variance (31, 73, 109, 121). Standard deviations for these studies were estimated from the pooled standard deviations of other studies reporting withdrawal at similar time points. Meta-analysis demonstrated a significant reduction in mean serum cholesterol of −0.39 mmol/L (random effects, CI −0.59 to −0.19, P<0.0001). However, heterogeneity was significant (I2=73.5%). Similar results are seen when only studies reporting intention-to-treat analysis are included (nine studies, random effects, WMD −0.33 mmol/L, CI −0.54 to −0.12, P=0.002, I2=62.2%). Sensitivity analysis removing those studies reporting no measure of variance does not affect the results and there is no interaction between study quality and effect size. The funnel plot is grossly symmetrical. Unlike the data for the risk of hypercholesterolemia, there is no interaction between the mean difference seen and the time of steroid withdrawal. Heterogeneity cannot be explained by concomitant immunosuppression.

Serum triglyceride levels are reported in 12 studies. Of these, only Woodle et al. (31) does not report a measure of variance, and so the standard deviation is estimated from the other studies in the steroid induction group. Meta-analysis demonstrates no significant reduction in triglyceride levels with SAW (random effects, WMD −0.17 mmol/L, CI −0.39 to −0.05, P=0.14). Heterogeneity is significant (I2=81.1%). Sensitivity and subgroup analyses fail to demonstrate a cause for this heterogeneity. The funnel plot is asymmetrical suggesting that there may be unpublished studies demonstrating larger reductions in triglycerides.

Other Outcomes

The effect of steroid avoidance and withdrawal on other outcomes of interest is shown in Table 3. Risk of infection and malignancy are unaffected, although the risk of leucopenia is significantly increased (fixed effects, RR 1.66, CI 1.42–1.93, P<0.0001). Overall, no difference in the risk of cataracts is observed, although there is significant heterogeneity in this analysis (I2=53.6%).

TABLE 3
TABLE 3:
Meta-analysis of other steroid-related side effects

A number of studies reported outcomes reflecting the effect of steroid therapy on bone metabolism. As the number of studies reporting such outcomes is small, and the reported outcomes vary, a narrative review is considered more appropriate than statistical meta-analysis.

Aroldi et al. (33) evaluated lumbar vertebral bone mineral density (BMD) in a subset of patients from the multicenter trial of Ponticelli et al. (67, 68). Study patients in whom steroids were withdrawn after the induction period demonstrated a significant increase in BMD during 18-month follow-up. In contrast, BMD decreased by 64% in patients in whom steroids were continued (P<0.001). These results are supported by the study by Vanrenterghem et al. (34), who demonstrated a significantly higher bone density in the L2/L3 vertebrae at 12 months in patients in whom steroids were withdrawn 3 months posttransplant. However, no difference was seen in bone density at the L4 vertebra or femoral neck.

Similar results were observed in a subset of patients in the study by Pelletier et al. (35, 36), in whom BMD in both the spine and hip increased in the steroid withdrawal group during the first year after withdrawal. No significant change was seen in the patients continuing steroids. This increase in BMD is seen even after late steroid withdrawal, with Farmer et al. (37) reporting significant increases in the BMD of the lumbar spine and femur in a group of patients undergoing steroid withdrawal greater than 1 year posttransplant.

Measurements of bone density are used in these trials as a surrogate marker for the risk of clinical events such as fractures. In a recent report from the ongoing trial from the Astellas Steroid Withdrawal Group, Woodle et al. (31) reported a significant reduction in the incidence of fractures and avascular necrosis at 5 years after steroid withdrawal (P=0.04).

DISCUSSION

The main concern with any immunosuppression withdrawal protocol after transplantation is that the risk of acute and chronic rejection will increase, potentially leading to graft damage and dysfunction or loss. The present meta-analysis in renal transplant recipients is based on 34 studies and a total of 5637 patients. It demonstrates an overall RR of AR of 1.56 with SAW. This finding is in keeping with previous meta-analyses, although the RRs in these previous studies have been greater (6–11). These previous meta-analyses have included smaller numbers of studies than the present analysis, often with selected subsets of patients on specific immunosuppressive regimens. Although such an increase in AR rate is of some concern, it should be noted that the baseline risk of AR with modern regimens is approximately 10% to 15% at 1 year and so the absolute increase in risk is relatively small in magnitude.

In keeping with the meta-analysis by Tan et al. (10), who reported that the majority of the increase seen in their study was mild AR of Banff grade I, there was no increase in corticosteroid resistant AR with SAW. Perhaps as a consequence of the mild rejection seen, overall graft survival does not differ significantly between groups. However, there is a trend toward an increase in death-censored graft loss with SAW (HR 1.19, P=0.19), which is stronger when only high-quality studies are included in the analysis (HR 1.31, P=0.07). This trend is lost when all graft losses (including deaths with a functioning graft) are considered (HR 0.99, P=0.93), which may result from a nonsignificant reduction in deaths in the avoidance or withdrawal arm (HR 0.82, P=0.2). Thus, it is possible that the increase in AR seen in the SAW arm slightly increases the risk of graft damage and loss, but this is offset because of a reduction in death from other causes (such as cardiovascular events) in these patients. The finding that the relatively mild AR episodes resulting from steroid avoidance and withdrawal do not have a large impact on long-term graft survival and function is supported by previous case series, in which survival has been related to severity of AR episodes (38, 39). Although there is a statistically significant decrease in graft function, with an increase in serum creatinine of 4.24 μmol/L and a decrease in creatinine clearance of 3.06 mL/min, such small changes are unlikely to be of great clinical significance.

Much of this efficacy data contrast with the findings from the prospective registry study of Opelz et al. (13) described earlier. Patients in this study were withdrawn from steroids at 6 months or later after transplantation, and compared with matched controls in the Collaborative Transplant Study Registry. In contrast to the outcomes identified here, no differences in AR rates or graft function were identified at 7 years after transplantation. In addition, graft and patient survivals were improved in those patients in whom steroids are withdrawn. There are a number of reasons why our current results and that from Opelz may conflict. The most likely explanation is the nonrandomized nature of the CTS study and the likelihood of significant selection bias, as has already been suggested (40). A recent commentary on this topic has suggested that as the effect sizes from interventions in transplantation are relatively small, randomization is essential to ensure that random errors and bias are significantly smaller than the effect size to be measured (41). A counter argument is that registry data allow for much larger numbers of patients to be analyzed, with longer follow-up times. Certainly, the 7-year follow-up in the study by Opelz is longer than the average for studies in the present meta-analysis, although the number of patients (2125 renal recipients) in the CTS study is in fact less than half that in this study (5637 recipients). Other differences in the CTS study are that steroids are withdrawn relatively late (>6 months posttransplantation) compared with those in our analysis. However, no association between withdrawal times and efficacy outcomes was identified in this meta-analysis. Nevertheless, there is no major difference in the general conclusions of the CTS data and this meta-analysis.

Of considerable importance is the finding that there is a significant reduction in cardiovascular risk factors. Hypertension, hypercholesterolemia, and new-onset diabetes are all reduced in patients in whom steroids are avoided or withdrawn. Cardiovascular disease is the most common cause of death with a functioning graft, and therefore any reduction in risk factors is likely to result in a long-term survival benefit (2). However, it is of note that a recent report of the impact of a reduction in some of these risk factors in dialysis patients, in contrast to a normal population, failed to show a reduction in cardiovascular events (42). Although this potential survival benefit is not seen in this analysis, the majority of studies have relatively short follow-up periods and small patient numbers. Individual studies cannot achieve enough statistical power to demonstrate differences in such uncommon long-term outcomes, and those studies included in this review that do report incidence of cardiovascular events unsurprisingly do not demonstrate any significant difference between groups. The only available method for demonstrating differences in these outcomes is likely to be analysis of long-term registry data. Unfortunately, these outcomes were not recorded in the study from Opelz et al., although secondary analysis of risk factors did demonstrate benefits in reduction in hypertension and hypercholesterolemia in patients in whom steroids were withdrawn (13).

Of interest is that later withdrawal of steroids seems to increase the reduction in the risk of hypercholesterolemia. The reason for this is unclear—one might expect the benefit of steroid withdrawal to be greatest when steroids are withdrawn early, thus reducing the total steroid exposure in the study group. It is possible that prolonged steroid use in the study group increases the risk of adrenal insufficiency on withdrawal, leading to a greater fall in serum lipids because of low endogenous steroid levels on steroid cessation. This hypothesis is not, however, supported by the data on measured serum cholesterol levels, which shows no relationship between the time of withdrawal and the reduction in serum cholesterol.

Neither the increase in AR nor the decrease in cardiovascular risk factors observed relate to the type of concurrent immunosuppression used. Perhaps more surprisingly, there also seems to be no relationship between withdrawal time and the effect size seen for outcomes other than serum cholesterol. It is possible that this is just a factor of reduced patient numbers in the subgroup analyses, particularly in the complete avoidance and late withdrawal groups, which reduces statistical power. For example, there is a suggestion that the risk of AR does not differ significantly between the withdrawal and maintenance groups when steroids are withdrawn beyond 12 months posttransplantation. However, the small number of studies in this group means that the RR is not significantly different from the other time points analyzed.

Another interesting finding is the increased risk of leucopenia with SAW. In normal individuals, steroids have a myeloproliferative effect, resulting in a leucocytosis. It is likely that when steroids are withdrawn, the myelosuppressive effect of other agents such as AZA or mycophenolate mofetil predominates and results in a relative leucopenia. This may be compounded by pharmacokinetic interactions between steroids and other immunosuppressants. Steroids are known to induce uridine diphosphate-glucuronosyltransferase activity in a time- and dose-dependent manner, which may increase the clearance of mycophenolic acid (43). Indeed, steroid tapering has been shown to result in an increase in the mycophenolic acid area under the curve (AUC), which may increase the risk of leucopenia (44). Corticosteroids are known to induce the cytochrome P450 3A4 isoenzyme, and this has been suggested as the mechanism for the increased tacrolimus levels seen after withdrawal of steroids (45–47).

Any benefits seen from steroid withdrawal or avoidance will be at least in part offset by the need for some patients to return to long-term steroids after AR episodes or deterioration in graft function. Humar et al. (48) suggested that addition of maintenance steroids in patients experiencing AR may help to reduce the risk of recurrent episodes. The overall benefit in a population treated in this way has been estimated in this analysis by only considering those studies in which an intention-to-treat analysis is reported. In these studies, patients are analyzed according to the randomization group regardless of whether steroids are added to maintenance immunosuppression after AR episodes. The present data in renal transplant recipients suggest that although, as expected, the benefits in the reduction of hypertension, hypercholesterolemia, and new-onset diabetes are reduced when only these studies are included, a significant overall benefit remains when compared with a patient population continuing steroids. This suggests that a population on an avoidance or withdrawal protocol will benefit overall, despite some patients returning to maintenance steroid therapy.

Although the inclusion criteria for the studies in this review are varied, the majority recruit only low-risk transplant recipients, with low or no rejection episodes before randomization and favorable tissue matching. Therefore, generalization of these results to other transplant populations may be inappropriate depending on underlying risk.

Weaknesses in this study are several. The overall methodological quality of the studies included ranges from good to poor. However, sensitivity analyses incorporating only high-quality studies did not affect any of the outcomes analyzed. In a number of outcomes, significant heterogeneity was identified, which could not be explained by predefined subgroup analyses for clinical and methodological factors such as concomitant immunosuppression, withdrawal time, or study quality. The lack of explanation could result from a lack of statistical power in subgroup analysis or from unidentified clinical or methodological factors resulting in heterogeneity. Although this unexplained heterogeneity is of concern, the use of a random-effects model in these cases incorporates the variance in effect size observed between studies, providing a more conservative estimate of the confidence interval for the overall effect. A further caveat to the present results is the suggestion of publication biases in a number of the funnel plots, in particular those for hypercholesterolemia, hypertension, and new-onset diabetes, which may lead to overestimation of the reduction in RR. It is of note that these outcomes are only reported in a small subset of studies, and the bias seen may reflect a “reporting bias” with a tendency to only report such outcomes if the study demonstrates a difference between the withdrawal and maintenance groups.

Definitions of reported outcomes are often variable or not clearly specified. For example, a significant proportion of studies do not report the definition of AR used to instigate therapy. This is also a problem with the definitions of cardiovascular risk factors. Some studies report use of antihypertensive or lipid-lowering drugs, and some report on the incidence of hypertension or hypercholesterolemia (but again with variable definitions). Few studies use the American Diabetes Association/World Health Organization (ADA/WHO) criteria for the definition of diabetes. The assumption made in this analysis is that the same underlying effects are being measured whichever method is used.

Firm conclusions from the survival data in this analysis are limited by the quality of reporting of these outcomes, with estimations for the HRs used in the majority of studies. The HR is used in meta-analysis as it incorporates not only the number of events observed but also withdrawals from the study and time-to-event data. The majority of studies do not report survival outcomes as HRs, meaning that these must be estimated from the log-rank P value from Kaplan-Meier analysis or from the raw event rates. The latter method loses any time-to-event or withdrawal data, approximating the RR of events. The use of estimated HRs in this way allows the full survival data to be used where possible, making the best use of the available data.

Overall, it is likely that in renal transplant recipients, the long-term benefits of a reduction in cardiovascular risk will outweigh the increased risk of AR with steroid-sparing or withdrawal protocols. The optimum time for withdrawal is less clear and although steroids may still be required to protect against the adverse effects of antibody induction therapies in the perioperative period, there is little evidence to suggest that there is any continuing benefit of steroid maintenance beyond this period in a low-risk population on modern immunosuppressive regimens.

REFERENCES

1.Starzl TE. Chapter Eighteen: Pretreatment with prednisolone. In: Experience in Renal Transplantation. Philadelphia, W.B. Saunders Company 1964.
2.Ojo AO, Hanson JA, Wolfe RA, et al. Long-term survival in renal transplant recipients with graft function. Kidney Int 2000; 57: 307.
3.Cheigh JS, Haschemeyer RH, Wang JC, et al. Hypertension in kidney transplant recipients. Effect on long-term renal allograft survival. Am J Hypertens 1989; 2(5 Pt 1): 341.
4.Dimeny E, Tufveson G, Lithell H, et al. The influence of pretransplant lipoprotein abnormalities on the early results of renal transplantation. Eur J Clin Invest 1993; 23: 572.
5.Veenstra DL, Best JH, Hornberger J, et al. Incidence and long-term cost of steroid-related side effects after renal transplantation. Am J Kidney Dis 1999; 33: 829.
6.Hricik DE, O'Toole MA, Schulak JA, et al. Steroid-free immunosuppression in cyclosporine-treated renal transplant recipients: A meta-analysis. J Am Soc Nephrol 1993; 4: 1300.
7.Kasiske BL, Chakkera HA, Louis TA, et al. A meta-analysis of immunosuppression withdrawal trials in renal transplantation. J Am Soc Nephrol 2000; 11: 1910.
8.Pascual J, Quereda C, Zamora J, et al. Steroid withdrawal in renal transplant patients on triple therapy with a calcineurin inhibitor and mycophenolate mofetil: A meta-analysis of randomized, controlled trials. Transplantation 2004; 78: 1548.
9.Pascual J, Quereda C, Zamora J, et al. Updated meta-analysis of steroid withdrawal in renal transplant patients on calcineurin inhibitor and mycophenolate mofetil. Transplant Proc 2005; 37: 3746.
10.Tan JY, Zhao N, Wu TX, et al. Safety and efficacy of steroid withdrawal in modern triple immunosuppressant: A systematic review [Chinese]. Chin J Evid Based Med 2006; 6: 273.
11.Tan JY, Zhao N, Wu TX, et al. Steroid withdrawal increases risk of acute rejection but reduces infection: A meta-analysis of 1681 cases in renal transplantation. Transplant Proc 2006; 38: 2054.
12.Opelz G. Effect of the maintenance immunosuppressive drug regimen on kidney transplant outcome. Transplantation 1994; 58: 443.
13.Opelz G, Dohler B, Laux G. Long-term prospective study of steroid withdrawal in kidney and heart transplant recipients. Am J Transplant 2005; 5(4 Pt 1): 720.
14.Knight SR. Optimisation of immunosuppressive drug therapy in transplant recipients. MChir Thesis, University of Cambridge 2009: 132.
15.Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Control Clin Trials 1996; 17: 1.
16.R Development Core Team. R: A language and environment for statistical computing: R Foundation for Statistical Computing, Vienna, Austria 2005.
17.Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21: 1539.
18.Vincenti F, Schena F, Walker R, et al. 3 months interim results of a 12-month study with enteric-coated mycophenolate sodium (EC-MPS, Myfortic®), basiliximab, and neoral C-2 comparing different steroid protocols in de novo kidney recipients [abstract]. J Am Soc Nephrol 2005; 16: 236a.
19.Walker R, Vincenti F, Schena FP, et al. Preliminary results of a 12-month study with enteric-coated mycophenolate sodium (EC-MPS), basiliximab, and neoral C-2 comparing two investigational steroid regimens (without steroids or short-term use of steroids) with standard steroid treatment in de novo kidney recipients [abstract]. Nephrology 2005; 10(suppl): A214.
20.Vincenti F, Schena FP, Paraskevas S, et al. Metabolic effects of steroid avoidance or early steroid withdrawal: 12-month results of a randomised trial in de novo renal transplant patients receiving cyclosporine, enteric-coated mycophenolate sodium (EC-MPS) and basiliximab [abstract]. Am J Transplant 2006; 6(suppl 2): 483.
21.Schena FP, Vincenti F, Paraskevas S, et al. 12-month results of a prospective, randomised trial of steroid avoidance, steroid withdrawal and standard steroids in de novo renal transplant patients receiving cyclosporine, enteric-coated mycophenolate sodium (EC-MPS) and basiliximab [abstract]. Am J Transplant 2006; 6(suppl 2): 84.
22.Vincenti F, Schena FP, Walker R, et al. Preliminary 3-month results comparing immunosuppressive regimens of enteric-coated mycophenolate sodium (EC- MPS) without steroids vs short-term use of steroids vs standard steroid treatment including basiloid treatment including basiliximab, and neoral C-2 in de novo kidney recipients abstract. Am J Transplant 2005; 5(11 suppl): 548.
23.Walker R, Chadban S, Russ G, et al. Comparison of metabolic parameters in renal transplant patients randomised to steroid avoidance, steroid withdrawal or standard steroids within a 12 month randomised multicentre trial [abstract]. Transplant Soc Aust N Z 2007: 32.
24.Vincenti F, Schena FP, Paraskevas S, et al. A randomized, multicenter study of steroid avoidance, early steroid withdrawal or standard steroid therapy in kidney transplant recipients. Am J Transplant 2008; 8: 307.
25.Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998; 17: 2815.
26.Walter SD, Yao X. Effect sizes can be calculated for studies reporting ranges for outcome variables in systematic reviews. J Clin Epidemiol 2007; 60: 849.
27.Furukawa TA, Barbui C, Cipriani A, et al. Imputing missing standard deviations in meta-analyses can provide accurate results. J Clin Epidemiol 2006; 59: 7.
28.Altman DG, Bland JM. Interaction revisited: The difference between two estimates. BMJ 2003; 326: 219.
29.Viechtbauer W. MiMa: An S-Plus/R function to fit meta-analytic mixed-, random-, and fixed-effects models [computer software and manual]. Available at: http://www.wvbauer.com/downloads.html. Accessed July 1, 2009.
30.Nematalla A, Bakr M, Gheith O, et al. Steroid avoidance immunosuppression: Long term evaluation in live donor renal allotransplant recipients [abstract]. 11th Annual Meeting of the British Transplantation Society, SECC, Glasgow, April 16–18, 2008.
31.Woodle ES, First MR, Pirsch J, et al. A prospective, randomized, double-blind, placebo-controlled multicenter trial comparing early (7 day) corticosteroid cessation versus long-term, low-dose corticosteroid therapy. Ann Surg 2008; 248: 564.
32.Kumar MS, Xiao SG, Fyfe B, et al. Steroid avoidance in renal transplantation using basiliximab induction, cyclosporine-based immunosuppression and protocol biopsies. Clin Transplant 2005; 19: 61.
33.Aroldi A, Tarantino A, Montagnino G, et al. Effects of three immunosuppressive regimens on vertebral bone density in renal transplant recipients: A prospective study. Transplantation 1997; 63: 380.
34.Vanrenterghem Y, Lebranchu Y, Hene R, et al. Double-blind comparison of two corticosteroid regimens plus mycophenolate mofetil and cyclosporine for prevention of acute renal allograft rejection. Transplantation 2000; 70: 1352.
35.Ing S, Sinnott L, Davies E, et al. Bone mineral density changes in a prospective randomized trial of steroid withdrawal in kidney transplant recipients abstract. Am J Transplant 2007; 7(suppl 2): 310.
36.Pelletier RP, Akin B, Ferguson RM. Prospective, randomized trial of steroid withdrawal in kidney recipients treated with mycophenolate mofetil and cyclosporine. Clin Transplant 2006; 20: 10.
37.Farmer CK, Hampson G, Abbs IC, et al. Late low-dose steroid withdrawal in renal transplant recipients increases bone formation and bone mineral density. Am J Transplant 2006; 6: 2929.
38.Humar A, Kerr S, Gillingham KJ, et al. Features of acute rejection that increase risk for chronic rejection. Transplantation 1999; 68: 1200.
39.Kasiske BL, Kalil RS, Lee HS, et al. Histopathologic findings associated with a chronic, progressive decline in renal allograft function. Kidney Int 1991; 40: 514.
40.Steiner RW. In defense of steroids. Am J Transplant 2005; 5: 2334.
41.Baigent C, Emberson JR. Randomization is essential for progress in transplant medicine. Transplantation 2008; 86: 26.
42.Fellstrom BC, Jardine AG, Schmieder RE, et al. Rosuvastatin and cardiovascular events in patients undergoing hemodialysis. N Engl J Med 2009; 360: 1395.
43.Schuetz EG, Hazelton GA, Hall J, et al. Induction of digitoxigenin monodigitoxoside UDP-glucuronosyltransferase activity by glucocorticoids and other inducers of cytochrome P-450p in primary monolayer cultures of adult rat hepatocytes and in human liver. J Biol Chem 1986; 261: 8270.
44.Cattaneo D, Perico N, Gaspari F, et al. Glucocorticoids interfere with mycophenolate mofetil bioavailability in kidney transplantation. Kidney Int 2002; 62: 1060.
45.Hesselink DA, Ngyuen H, Wabbijn M, et al. Tacrolimus dose requirement in renal transplant recipients is significantly higher when used in combination with corticosteroids. Br J Clin Pharmacol 2003; 56: 327.
46.Kim JS, Aviles DH, Silverstein DM, et al. Effect of age, ethnicity, and glucocorticoid use on tacrolimus pharmacokinetics in pediatric renal transplant patients. Pediatr Transplant 2005; 9: 162.
47.van Duijnhoven EM, Boots JM, Christiaans MH, et al. Increase in tacrolimus trough levels after steroid withdrawal. Transplant Int 2003; 16: 721.
48.Humar A, Gillingham K, Kandaswamy R, et al. Steroid avoidance regimens: A comparison of outcomes with maintenance steroids versus continued steroid avoidance in recipients having an acute rejection episode. Am J Transplant 2007; 7: 1948.
49.MacDonald AS, Daloze P, Dandavino R, et al. A randomized study of cyclosporine with and without prednisone in renal allograft recipients. Canadian Transplant Group. Transplant Proc 1987; 19(1 Pt 3): 1865.
50.Stiller C. The requirements for maintenance steroids in cyclosporine-treated renal transplant recipients. Transplant Proc 1983; 15(4 suppl 1–2): 2490.
51.Morales JM, Marcen R, Grino JM, et al. Comparison of three induction protocols in older renal transplant patients: Cyclosporine (CyA) monotherapy (M) vs OKT3 and CyA vs prednisone (P) and CyA. A multicenter randomized study [abstract]. J Am Soc Nephrol 1993; 4(Program and Abstracts): 950.
    52.Spanish Monotherapy Study Group. Cyclosporine monotherapy versus OKT3 and cyclosporine versus prednisone and cyclosporine as induction therapy in older renal transplant patients: A multicenter randomized study. Spanish Monotherapy Study Group. Transplant Proc 1994; 26: 2522.
    53.De Vecchi A, Tarantino A, Montagnino G, et al. Ciclosporin alone or combined with steroid in the treatment of cadaveric renal transplant recipients. Clin Transplant 1987; 1: 198.
      54.De Vecchi A, Tarantino A, Rivolta E, et al. Ciclosporin alone or associated with steroid for immunosuppression of cadaveric renal transplants? Contrib Nephrol 1986; 51: 88.
      55.De Vecchi A, Tarantino A, Rivoltan E, et al. Need for steroid in cyclosporine (Cy) treated cadaveric renal transplant recipients (pts) [abstract]. Kidney Int 1985; 28: 394.
      56.Griffin PJ, Gomes Da Costa CA, Salaman JR. Renal transplantation without steroids: A controlled clinical trial. Transplant Proc 1986; 18: 797.
      57.Griffin PJ, Gomes Da Costa CA, Salaman JR. A controlled trial of steroids in cyclosporine-treated renal transplant recipients. Transplantation 1987; 43: 505.
      58.Nott D, Griffin PJ, Salaman JR. Low-dose steroids do not augment cyclosporine immunosuppression but do diminish cyclosporine nephrotoxicity. Transplant Proc 1985; 17: 1289.
      59.Salaman JR, Gomes Da Costa CA, Griffin PJ. Renal transplantation without steroids. J Pediatr 1987; 111(6 Pt 2): 1026.
      60.Johnson RW, Mallick NP, Bakran A, et al. Cadaver renal transplantation without maintenance steroids. Transplant Proc 1989; 21(1 Pt 2): 1581.
      61.Johnson RW, Mallick NP, Scott PD, et al. Prospective trials with cyclosporine monotherapy in cadaver renal transplantation. J Nephrol 1990; 3(4 suppl 1): 47.
        62.Hricik DE, Moritz C, Mayes JT, et al. Association of the absence of steroid therapy with increased cyclosporine blood levels in renal transplant recipients. Transplantation 1990; 49: 221.
        63.Schulak JA, Mayes JT, Moritz CE, et al. A prospective randomized trial of prednisone versus no prednisone maintenance therapy in cyclosporine-treated and azathioprine-treated renal transplant patients. Transplantation 1990; 49: 327.
        64.Schulak JA, Moritz CE, Hricik DE. Renal transplantation without prednisolone: Effects of bone marrow tolerance azathioprine. Transplant Proc 1989; 21(1 Pt 2): 1709.
        65.Montagnino G, Tarantino A, Segoloni GP, et al. Long-term results of a randomized study comparing three immunosuppressive schedules with cyclosporine in cadaveric kidney transplantation. J Am Soc Nephrol 2001; 12: 2163.
        66.Ponticelli C, Tarantino A, Montagnino G. Steroid withdrawal in renal transplant recipients. Transplant Proc 2001; 33: 987.
        67.Ponticelli C, Tarantino A, Segoloni GP, et al. A randomized study comparing three cyclosporine-based regimens in cadaveric renal transplantation. Italian Multicentre Study Group for Renal Transplantation (SIMTRe). J Am Soc Nephrol 1997; 8: 638.
        68.Ponticelli C, Tarantino A, Segoloni GP, et al. A randomized study comparing cyclosporine alone vs double and triple therapy in renal transplants. The Italian Multicentre Study Group for Renal Transplantation (SIMTRe). Transplant Proc 1997; 29: 290.
        69.Tarantino A, Italian Multicentre Study Group for Renal T. Is cylosporine (CsA, sandimmun) monotherapy an effective and safe immunosuppressant in renal transplant recipients? [abstract]. 16th Annual Meeting. American Society of Transplant Physicians (ASTP); May 10–14, 1997; Chicago (ILL) 1997: 128.
          70.Tarantino A, per il Gruppo SIMTRE. Long-term results of a randomized trial of 3 cyclosporine (CSA) regimens in cadaveric kidney allografts (TXC) [abstract]. Nephrol Dial Transplant 2000; 15: 250.
            71.Tarantino A, Segoloni GP, Cambi V, et al. A randomized study comparing three cyclosporine-based regimens in cadaveric renal transplantation: Results at 7 years. Transplant Proc 1998; 30: 1729.
            72.Painter PL, Topp KS, Krasnoff JB, et al. Health-related fitness and quality of life following steroid withdrawal in renal transplant recipients. Kidney Int 2003; 63: 2309.
            73.Vincenti F, Monaco A, Grinyo J, et al. Multicenter randomized prospective trial of steroid withdrawal in renal transplant recipients receiving basiliximab, cyclosporine microemulsion and mycophenolate mofetil. Am J Transplant 2003; 3: 306.
            74.Vincenti F, Monaco A, Grinyo J, et al. Rapid steroid withdrawal versus standard steroid treatment in patients treated with simulect, neoral, and cellcept for the prevention of acute rejection in renal transplantation: A multicenter, randomized trial [abstract]. Transplantation 2000; 69(8 suppl): S133.
            75.Klinger M, Vitko S, Salmela K, et al; Group ATLASS. Large, prospective study evaluating steroid-free immunosuppression with tacrolimus/basiliximab and tacrolimus/MMF compared with tacrolimus/MMF/steroids in renal transplantation [abstract]. Nephrol Dial Transplant 2003; 18(suppl 4): 788.
            76.Kramer BK, Klinger M, Salmela K, et al. Two steroid-free immunosuppressive regimens (basiliximab/tacrolimus and tacrolimus/MMF) in comparison to tacrolimus/MMF/steroid therapy after renal transplantation [abstract]. J Am Soc Nephrol 2003; 14: 9a.
            77.Kramer BK, Kruger B, Hoffmann U, et al. 1-year follow-up of two steroid-free immunosuppressive regimens—Basiliximab/Tacrolimus and tacrolimus/MMF—In comparison to tacrolimus/MMF/steroids after renal transplantation [abstract]. J Am Soc Nephrol 2004; 15: 289a.
              78.Vitko S, Klinger M, Salmela K, et al. Two corticosteroid-free regimens—Tacrolimus monotherapy after basiliximab administration and tacrolimus/mycophenolate mofetil—In comparison with a standard triple regimen in renal transplantation: Results of the Atlas Study. Transplantation 2005; 80: 1734.
              79.Vitko S, Klinger M, Salmela KW, et al; Group ATLASs. Comparison of two steroid-free regimens—Basiliximab/Tacrolimus and Tacrolimus/MMF—With Tacrolimus/MMF/Steroid Therapy After Renal Transplantation [abstract]. Am J Transplant 2003; 3(suppl 5): 312.
              80.Laftavi M, Stefanick B, Stephan R, et al. The significance of protocol biopsy in immunominimization protocol: A prospective study of steroid withdrawal [abstract]. Transplantation 2004; 78 (2 suppl): 89.
              81.Laftavi MR, Stephan R, Stefanick B, et al. Randomized prospective trial of early steroid withdrawal compared with low-dose steroids in renal transplant recipients using serial protocol biopsies to assess efficacy and safety. Surgery 2005; 137: 364.
              82.Pankewycz OG, Stephan R, Stefanick B, et al. The clinical benefits of early steroid withdrawal (7 days) and utility of protocol biopsies at 1, 6 and 12 months in guiding steroid-free immunosuppressive therapy after renal transplantation [abstract]. Am J Transplant 2004; 4(suppl 8): 579.
              83.Pankewyez O, Stephan R, Stefanick B, et al. Induction immunosuppression for renal transplantation using thymoglobulin, FK506 and mycophenolate mofetil allows for safe steroid withdrawal and eliminates the need for early protocol biopsies [abstract]. Am J Transplant 2003; 3(suppl 5): 478.
              84.Montagnino G, Sandrini S, Casciani C, et al. A randomized trial of steroid avoidance in renal transplant patients treated with everolimus and cyclosporine. Transplant Proc 2005; 37: 788.
              85.Montagnino G, Sandrini S, Iorio B, et al. A randomized exploratory trial of steroid avoidance in renal transplant patients treated with everolimus and low-dose cyclosporine. Nephrol Dial Transplant 2008; 23: 707.
              86.Ponticelli C, Sandrini S, Casciani C, et al; Group STAR. A randomized trial of steroid avoidance in renal transplant patients treated with everolimus and cyclosporine [abstract]. Transplantation 2004; 78 (2 suppl): 170.
              87.Gaber AO, Moore LW, Pirsch J, et al. Characteristics of rejection in renal allografts following early corticosteroid withdrawal in a randomized controlled clinical trial: Results of a 3 year follow-up [abstract]. Transplantation 2006; 82(1 suppl 3): 178.
              88.Pirsch J, Woodle ES, Shihab F, et al. Effect of steroid withdrawal on new onset diabetes after transplant: Results of a randomized double blinded placebo controlled trial [abstract]. Transplantation 2006; 82(1 suppl 3): 355.
              89.Woodle ES. A randomized double blind, placebo-controlled trial of early corticosteroid cessation versus chronic corticosteroids: Four year results [abstract]. Am J Transplant 2007; 7(suppl 2).
              90.Woodle ES, Gaber AO, Shihab F, et al. Comparison of T cell depleting and non-t cell depleting antibody induction therapy for early corticosteroid withdrawal regimens abstract. Transplantation 2006; 82(1 suppl 3): 353.
              91.Woodle ES. A randomsied double blind, placebo-controlled trial of early corticosteroid cessation versus chronic corticosteroids: Five year results. Am J Transplant 2008; 8(2 suppl): 300.
              92.Nematalla AH, Bakr M, Elagroudy A, et al. Improving quality of life after steroid avoidance immunosuppression regimen in live donor renal allotransplant recipients—A prospective randomized controlled study single center experience (Two year follow up) [abstract]. Transpl Int 2007; 20(suppl 2): 134.
              93.Nematalla AH, Bakr MA, Gheith OA, et al. Steroid-avoidance immunosuppression regimen in live-donor renal allotransplant recipients: A prospective, randomized, controlled study. Exp Clin Transplant 2007; 5: 673.
                94.Cristinelli L, Brunori G, Manganoni AM, et al. Controlled study of steroid withdrawal after 6 months in renal transplant patients treated with ciclosporin. Contrib Nephrol 1986; 51: 91.
                95.Cristinelli L, Brunori G, Setti G, et al. Withdrawal of methylprednisolone at the sixth month in renal transplant recipients treated with cyclosporine. Transplant Proc 1987; 19(1 Pt 3): 2021.
                96.Maiorca R, Cristinelli L, Brunori G, et al. Prospective controlled trial of steroid withdrawal after six months in renal transplant patients treated with cyclosporine. Transplant Proc 1988; 20(3 suppl 3): 121.
                97.Isoniemi H. Renal allograft immunosuppression. III. Triple therapy versus three different combinations of double drug treatment: Two year results in kidney transplant patients. Transpl Int 1991; 4: 31.
                98.Isoniemi H. Renal allograft immunosuppression. V. Glucose intolerance occurring in different immunosuppressive treatments. Clin Transplant 1991; 5: 268.
                  99.Isoniemi H, Ahonen J, Eklund B, et al. Renal allograft immunosuppression. II. A randomized trial of withdrawal of one drug in triple drug immunosuppression. Transplant Int 1990; 3: 121.
                  100.Isoniemi H, Eklund B, Hockerstedt K, et al. Discontinuation of one drug in triple drug treatment of renal allograft patients: 1-year results. Transplant Proc 1990; 22: 1365.
                  101.Isoniemi H, Tikkanen M, Häyry P, et al. Lipid profiles with triple drug immunosuppressive therapy and with double drug combinations after renal transplantation and stable graft function. Transplant Proc 1991; 23(1 Pt 2): 1029.
                  102.Isoniemi H, Tikkanen MJ, Ahonen J, et al. Renal allograft immunosuppression. IV. Comparison of lipid and lipoprotein profiles in blood using double and triple immunosuppressive drug combinations. Transplant Int 1991; 4: 130.
                  103.Isoniemi HM, Ahonen J, Tikkanen MJ, et al. Long-term consequences of different immunosuppressive regimens for renal allografts. Transplantation 1993; 55: 494.
                  104.Sinclair NR. Low-dose steroid therapy in cyclosporine-treated renal transplant recipients with well-functioning grafts. The Canadian Multicentre Transplant Study Group. Can Med Assoc J 1992; 147: 645.
                  105.Gulanikar AC, Belitsky P, MacDonald AS, et al. Randomized controlled trial of steroids versus no steroids in stable cyclosporine-treated renal graft recipients. Transplant Proc 1991; 23(1 Pt 2): 990.
                  106.Kim HC, Chang KJ, Kwon JK, et al. Long-term results of cyclosporine monotherapy in renal transplantation. Transplant Proc 1998; 30: 3539.
                  107.Park K, Kim ST, Lee SR, et al. A 1-year prospective randomized study in Korean living donor kidney transplant recipients: Comparing cyclosporine monotherapy and cyclosporine/prednisolone during the maintenance phase of immunosuppression. Transplant Proc 1994; 26: 1985.
                  108.Aswad S, Zapanta R, Wu L, et al. Steroid withdrawal in living related kidney transplant patients receiving FK506 [abstract]. Nephrol Dial Transplant 1998; 13: A285.
                  109.Ahsan N, Hricik D, Matas A, et al. Prednisone withdrawal in kidney transplant recipients on cyclosporine and mycophenolate mofetil—A prospective randomized study. Steroid Withdrawal Study Group. Transplantation 1999; 68: 1865.
                  110.Matas A, Ewell M. Prednisone withdrawal in kidney transplant recipients on CSA/MMF—A prospective randomized study abstract. Transplantation 1999; 67: S543.
                  111.Hene R; for the M55002 Group. A randomized, double-blind multicenter trial comparing two corticosteroid regimens in combination with mycophenolate mofetil (MMF) and cyclosporine (CYA) in renal transplant recipients abstract. Transplantation 1998; 65: S107.
                  112.Lebranchu Y. Comparison of two corticosteroid regimens in combination with CellCept and cyclosporine A for prevention of acute allograft rejection: 12 month results of a double-blind, randomized, multi- center study. M 55002 Study Group. Transplant Proc 1999; 31: 249.
                  113.Lebranchu Y, Aubert P, Bayle F, et al. Could steroids be withdrawn in renal transplant patients sequentially treated with ATG, cyclosporine, and cellcept? One-year results of a double-blind, randomized, multicenter study comparing normal dose versus low-dose and withdrawal of steroids. Transplant Proc 2000; 32: 396.
                  114.Lebranchu Y; on behalf of M55002. A comparison of two corticosteroid regimens in kidney transplanted patients treated with ATG/OKT3 induction, mycophenolate mofetil (MMF) and cyclosporine A (CyA) for prevention of acute allograft rejection. 12 months of a double-blind, randomized, multi-center study [abstract]. Transplantation 1999; 67: S239.
                  115.Nowacka-Cieciura E, Durlik M, Cieciura T, et al. Elevated serum immunoglobulins after steroid withdrawal in renal allograft recipients. Transplant Proc 2002; 34: 564.
                  116.Nowacka-Cieciura E, Durlik M, Cieciura T, et al. Steroid withdrawal after renal transplantation—Risks and benefits. Transplant Proc 2002; 34: 560.
                  117.Nowacka-Cieciura E, Durlik M, Cieciura T, et al. Positive effect of steroid withdrawal on bone mineral density in renal allograft recipients. Transplant Proc 2001; 33: 1273.
                  118.Nowacka-Cieciura E, Soluch L, Cieciura T, et al. Effect of glucocorticoid-free immunosuppressive protocol on serum lipids in renal transplant patients. Transplant Proc 2000; 32: 1339.
                  119.Puig i Marí JM. Induction treatment with mycophenolate mofetil, cyclosporine, and low-dose steroids with subsequent early withdrawal in renal transplant patients: Results of the Spanish Group of the Cellcept study. Transplant Proc 1999; 31: 2256.
                  120.Pisani F, Buonomo O, Iaria G, et al. Preliminary results of a prospective randomized study of basiliximab in kidney transplantation. Transplant Proc 2001; 33: 2032.
                  121.Burke J, Francos BB, Francos GC. Double-blind, placebo-controlled trial of steroid withdrawal in kidney transplant recipients with a cyclosporine/mycophenolate regimen-three year follow up [abstract]. Am J Transplant 2001; 1(suppl 1): 296.
                  122.Francos GC, Frankel CJ, Dunn SR, et al. Double-blind, placebo-controlled, 3 year study of steroid withdrawal using a neoral and mycophenylate mofetil (MMF)-based immunosuppressive regimen in primary renal transplant recipients [abstract]. Am J Transplant 2002; 2(suppl 3): 172.
                  123.Boletis JN, Konstadinidou I, Chelioti H, et al. Successful withdrawal of steroid after renal transplantation. Transplant Proc 2001; 33: 1231.
                  124.Boletis JN, Konstadinidou I, Darema M, et al. Steroids withdrawal in renal transplant recipients: A randomized controlled study [abstract]. Nephrol Dial Transplant 2002; 17(suppl 7): 312.
                  125.de Sevaux RG, Smak Gregoor PJ, Hene R, et al. Withdrawal of cyclosporine or prednisone in renal transplant recipients treated with mycophenolate mofetil, cyclosporine, and prednisone: A randomised study [abstract]. Transplantation 1999; 67: S240.
                  126.Smak Gregoor PJ, De Sevaux RG, Ligtenberg G, et al. A prospective randomised study of withdrawal of cyclosporine or prednisone in renal transplant recipients treated with mycophenolate mofetil, cyclosporine, and prednisone: 18 months follow-up data [abstract]. Am J Transplant 2001; 1(suppl 1): 246.
                  127.Smak Gregoor PJ, de Sévaux RG, Ligtenberg G, et al. Withdrawal of cyclosporine or prednisone six months after kidney transplantation in patients on triple drug therapy: A randomized, prospective, multicenter study. J Am Soc Nephrol 2002; 13: 1365.
                  128.Smak Gregoor PJH, van Gelder T, Ijzermans JNM, et al. Long-term results of a randomized, prospective study after withdrawal of cyclosporine or prednisone in renal transplant recipients treated with mycophenolate mofetil, cyclosporine, and prednisone [abstract]. Am J Transplant 2003; 3(suppl 5): 217.
                  129.Van Gelder T, De Sevaux R, Hene R, et al. Discontinuation of cyclosporine or prednisone 6 months after kidney transplantation: A randomized trial [abstract]. J Am Soc Nephrol 2001; 12(Program and Abstracts): 920a.
                    130.Sola E, Alferez MJ, Cabello M, et al. Low-dose and rapid steroid withdrawal in renal transplant patients treated with tacrolimus and mycophenolate mofetil. Transplant Proc 2002; 34: 1689.
                    131.Budde K, Salmela K, Pascual J, et al; Group THOMASFuS. Steroid-withdrawal in tacrolimus-treated renal transplant recipients: Results of a 3-year follow-up study [abstract]. Third International Congress on Immunosuppression, San Diego, CA, December 8–11, 2004.
                      132.Jindal RM, Salmela K, Vanrenterghem Y, van Hooff J, Squifflet JP. Reduction of high cholesterol levels by early withdrawal of steroids from a tacrolimus-based triple regimen [abstract]. Am J Transplant 2002; 2(S3): 190.
                      133.Pascual J, van Hooff JP, Salmela K, et al. Long-term efficacy and safety of steroid withdrawal in tacrolimus-treated renal transplant recipients: Results of a 3 year followup. Am J Transplant 2004; 4(suppl 8): 578.
                      134.Pascual J, van Hooff JP, Salmela K, et al. Three-year observational follow-up of a multicenter, randomized trial on tacrolimus-based therapy with withdrawal of steroids or mycophenolate mofetil after renal transplant. Transplantation 2006; 82: 55.
                      135.Rigotti P; European Tacrolimus MMFTSG. Patients with high cholesterol levels benefit most from early withdrawal of corticosteroids. Transplant Proc 2002; 34: 1797.
                      136.Squifflet JP, Vanrenterghem Y, van Hooff JP, et al., European Tacrolimus MMFTSG. Safe withdrawal of corticosteroids or mycophenolate mofetil: Results of a large, prospective, multicenter, randomized study. Transplant Proc 2002; 34: 1584.
                      137.Van den Ham ECH, Kooman JP, Christiaans MHL, et al. The influence of early steroid withdrawal on body composition and bone mineral density in renal transplantation patients. Transplant Int 2003; 16: 82.
                      138.van Hooff JP, European Tacrolimus MMFTSG. Effect of controlled steroid withdrawal on glucose levels in a tacrolimus-based immunosuppression regimen [abstract]. Third International Congress on Immunosuppression, San Diego, CA, December 8–11, 2004.
                        139.van Hooff JP, Vanrenterghem Y, Squifflet JP, et al; European Tacrolimus MMFTSG. First, large, prospective study of a controlled withdrawal of steroids or MMF following three months of tacrolimus/MMF/steroid therapy [abstract]. J Am Soc Nephrol 2001; 12(Program & Abstracts): 920a.
                        140.Vanrenterghem Y, van Hooff JP, Squifflet JP, et al. Minimization of immunosuppressive therapy after renal transplantation: Results of a randomized controlled trial. Am J Transplant 2005; 5: 87.
                        141.Dudley CR, Ratcliffe PJ, Higgins RM, et al. Effect of steroid withdrawal on graft function in renal transplant recipients [abstract]. Nephrol Dial Transplant 1994; 9: 1672.
                          142.Ratcliffe PJ, Dudley CR, Higgins RM, et al. Randomised controlled trial of steroid withdrawal in renal transplant recipients receiving triple immunosuppression. Lancet 1996; 348: 643.
                          143.Ratcliffe PJ, Firth JD, Higgins RM, et al. Randomized controlled trial of complete steroid withdrawal in renal transplant patients receiving triple immunosuppression. Transplant Proc 1993; 25(1 Pt 1): 590.
                          144.Hollander AA, Hene RJ, Van Es LA, et al. Late prednisone withdrawal in renal transplant patients [abstract]. J Am Soc Nephrol 1996; 7: 1911.
                          145.Hollander AA, Hene RJ, Hermans J, et al. Late prednisone withdrawal in cyclosporine-treated kidney transplant patients: A randomized study. J Am Soc Nephrol 1997; 8: 294.
                          146.Matl I, Lacha J, Lodererova A, et al. Withdrawal of steroids from triple-drug therapy in kidney transplant patients. Nephrol Dial Transplant 2000; 15: 1041.
                          147.Matl I, Lacha J, Lodererova A, et al. Withdrawal of steroids from the triple-drug therapy in renal transplant patients. Casopis Lekaru Ceskych 2000; 139: 115.
                          148.Matl I, Lacha J, Lodererova A, et al. Withdrawal of steroids from triple drug therapy in renal transplant patients [abstract]. Nephrol Dial Transplant 1998; 13: A256.
                          149.Kacar S, Gurkan A, Karaoglan M, et al. Steroid withdrawal protocol in renal transplantation [abstract]. 41st Congress. European Renal Association. European Dialysis and Transplantation Association, Lisbon, Portugal, May 15–18, 2004: 401.
                            150.Farmer C, Abbs I, Hilton R, et al. What is the value of short synacthen tests in predicting the ease of steroid withdrawal in renal transplant recipients? A randomised controlled trial [abstract]. Am J Transplant 2001; 1(suppl 1): 190.
                            151.Akin B, Ferguson RM, Pelletier RP. Five year follow-up after steroid withdrawal demonstrates no evidence of worsening renal function [abstract]. Am J Transplant 2004; 4(suppl 8): 298.
                            152.Pelletier RP, Davies EA, Elkhammas EA, et al. Randomized, prospective trial of prednisone withdrawal in stable renal transplant recipients [abstract]. Transplantation 2000; 69(8 suppl): S260.
                            Keywords:

                            Meta-analysis; Renal transplant; Steroid minimisation; Cardiovascular risk

                            © 2010 Lippincott Williams & Wilkins, Inc.