- Top of page
Anemia and erythrocytosis (PTE) are common after kidney transplantation. We sought to determine the influence of sirolimus compared to mycophenolate mofetil (MMF) on post-transplant erythropoiesis.
A total of 214 patients with recent kidney or kidney–pancreas transplants were treated with either sirolimus-based (n = 87) or MMF-based (n = 127) therapy. At 12 months, the prevalence of anemia was 31% with MMF and 57% with sirolimus (p < 0.001). Linear regression was used to examine the independent influence of sirolimus on hemoglobin at 12 months, controlling for multiple factors including gender and renal function. Sirolimus remained a significant correlate of lower hemoglobin in all patients (slope =−1.060, 95% CI: −1.76 to −0.362, p = 0.003), and in patients without PTE (slope =−0.671, 95% CI: −1.32 to −0.028, p = 0.041).
PTE, defined as a persistent hematocrit above 51%, occurred in 19% with MMF and 7% with sirolimus (p = 0.013). PTE was examined using logistic regression analysis. Sirolimus use correlated negatively with PTE (odds ratio with sirolimus = 0.33, 95% CI: 0.12 to 0.89, p = 0.028).
Our results indicate that, compared to treatment with MMF, treatment of kidney or kidney–pancreas recipients with sirolimus is associated with a higher prevalence of anemia, lower hemoglobin levels and lower incidence of PTE.
- Top of page
Anemia is common following kidney transplantation and has been recognized increasingly as a late complication of transplantation as graft survival has improved (1–3). Post-transplant anemia has been linked to multiple factors including poor allograft function, acute and chronic rejection, iron deficiency, viral infections, hemolytic uremic syndrome, treatment with angiotensin converting-enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) and immunosuppressive drugs (1–4). In patients with chronic kidney disease who are not yet on dialysis, correction of anemia has been shown to reduce left ventricular mass index (5). Anemia has been independently associated with the left ventricular hypertrophy, left ventricular dysfunction and congestive heart failure in patients on chronic dialysis (6,7). It is likely that these relationships between anemia and cardiovascular disease are also relevant in kidney transplant recipients. Indeed, anemia has been associated with de novo congestive heart failure and left ventricular hypertrophy after kidney transplantation (8,9).
Post-transplant erythrocytosis (PTE) is also common after kidney transplantation, occurring in 10–15% of patients (10). Risk factors for PTE include male gender, retention of native kidneys, good allograft function in the absence of rejection and high baseline hemoglobin pre-transplant (11,12). Erythrocytosis has been associated with venous and arterial thrombosis, myocardial infarction and stroke (11,13,14). ACEIs and ARBs are effective treatments for PTE, likely related to their influence on angiotensin II type I receptors, which are increased on erythroid progenitors in patients with PTE (15).
Some commonly used immunosuppressive drugs also may impact negatively on erythropoiesis. Azathioprine, a purine analog, can cause bone marrow suppression typically manifested by leukopenia and thrombocytopenia. Anemia and idiosyncratic pure red cell aplasia also are reported side effects of this agent (16,17). Mycophenolate mofetil (MMF) acts to block de novo purine synthesis and has been associated with generalized bone marrow suppression. Dose-dependent anemia associated with MMF has been reported in a large randomized prospective trial (18). A recent European survey also found that treatment with either azathioprine or MMF was an independent correlate of post-transplant anemia (19).
Sirolimus, a macrolide that inhibits the mammalian target of rapamycin (mTOR), is a potent inhibitor of cell cycle progression and not uncommonly causes thrombocytopenia and/or leukopenia. Anemia was a significant adverse event in a phase III trial in which sirolimus was administered in combination with cyclosporine and corticosteroids (20). Whether sirolimus has a greater impact on anemia and erythropoiesis than other antiproliferative agents is not clear. In the study reported herein, we compared the impact of erythropoiesis in two concurrent cohorts of patients receiving either sirolimus- or MMF-based immunosuppression.
- Top of page
A total of 314 patients received kidney or kidney–pancreas transplants between January 1999 and June 2002. Of these, 17 did not receive either MMF or sirolimus and were excluded from the analysis. We also excluded 51 patients who died (n = 14), lost their allografts (n = 18) or were lost to follow-up (n = 19) in the first year. Finally, we excluded 32 patients who were switched from MMF (n = 20) or sirolimus (n = 12) for a variety of reasons. In no patient was either MMF or sirolimus discontinued because of anemia.
After exclusions, 214 patients (127 on MMF-based therapy and 87 on sirolimus-based therapy) were included in the analysis. Clinical characteristics of patients in the two groups are compared in Table 1. The sirolimus-treated group contained more women and African Americans. Patients in the sirolimus group were more often treated with tacrolimus than with cyclosporine. All patients were treated initially with corticosteroids. However, steroids were withdrawn from 37 patients in the sirolimus group and in only 1 patient in the MMF group. The incidence of biopsy-proven acute rejection was higher in the sirolimus group (21% vs. 8%, p = 0.006). Serum creatinine concentrations were higher in the sirolimus group at both 6 and 12 months. At 12 months, estimated glomerular filtration rate was also significantly lower in the sirolimus group (63 ± 23 vs. 75 ± 25 mL/min, p < 0.001).
Table 1. Baseline data for MMF versus sirolimus treated patients
| ||MMF (n=127)||Sirolimus (n=87)||p Value |
|Female||44 (35%)||45 (52%)|| 0.013|
|African American||18 (14%)||65 (75%)||<0.001|
| Diabetes||62 (48%)||29 (33%)|| 0.024|
| Chronic GN||21 (17%)||15 (17%)||NS|
| Hypertension|| 7 (6%)||23 (26%)||<0.001|
| Other||32 (25%)||12 (14%)|| 0.043|
| Unknown|| 5 (4%)|| 8 (9%)||NS|
|Living donor||42 (33%)||30 (34%)||NS|
|Kidney–pancreas||22 (17%)|| 3 (3%)|| 0.002|
|Cyclosporine||24 (19%)|| 1 (1%)||<0.001|
|Steroid withdrawal||1 (0.8%)||37 (43%)||<0.001|
|Rejection||10 (8%)||18 (21%)|| 0.006|
|Infection||27 (21%)||15 (17%)||NS|
| Six months||1.4±0.4||1.8±0.8||<0.001|
| Twelve months||1.5±0.5||1.8±0.7||<0.001|
| Six months||74±24||65±24|| 0.014|
| Twelve months||75±25||63±23||<0.001|
The prevalence of anemia was higher in the sirolimus group than in the MMF group at both 6 months (57% vs. 41%, p = 0.024), and 12 months (57% vs. 31%, p < 0.001). The decrease in prevalence of anemia between 6 and 12 months in the MMF group was significant (p = 0.01), while the prevalence of anemia was unchanged over this time interval in the sirolimus group. Hb concentrations are depicted in Figure 1. At 6 months Hb averaged 13.3 ± 2.2 g/dL in the MMF group and 12 ± 1.9 g/dL in the sirolimus group (p < 0.001). Similarly, at 12 months Hb averaged 13.6 ± 2 g/dL and 12.1 ± 2 g/dL in the MMF and sirolimus groups, respectively (p < 0.001). There was no difference in the percentage of patients who received blood transfusions between groups over 12 months time (7.9% vs. 6.9% for MMF and sirolimus patients, respectively, p = NS).
Figure 1. Six and 12 month Hb levels (g/dL) for patients on sirolimus (n = 87) and MMF (n = 127). Shown are median values with interquartile range (box borders) and extreme values (whiskers).
Download figure to PowerPoint
Univariate analyses of variables associated with 12 month Hb are shown in Table 2. Female gender, African American ethnicity, kidney rather than kidney–pancreas transplant, biopsy-proven acute rejection, history of chronic infection, older recipient age, older donor age, higher serum creatinine concentration, lower creatinine clearance and use of sirolimus rather than MMF were all associated with lower Hb. Type of pre-existing kidney disease, living versus cadaveric donor source, cyclosporine versus tacrolimus and withdrawal of steroids were not associated with Hb in this analysis.
Table 2. Variables associated with 12 month Hb by univariate analysis
| Variable||Difference, mean Hb||95% CI of difference|| p Value|
|Female gender||−1.6||−2.13 to −1.04||<0.001|
|African American||−1.1||−1.66 to −0.51||<0.001|
|Kidney–pancreas|| 1.3|| 0.37 to 2.14|| 0.006|
|Sirolimus||−1.4||−1.97 to −0.86||<0.001|
|Rejection||−1.2||−2.05 to −0.37|| 0.005|
|Infection||−1.0||−1.70 to −0.27|| 0.007|
|Variable||Slope||95% CI of slope||p Value|
|Older age||−0.033||−0.06 to −0.01||<0.001|
|Older donor age||−0.029||−0.05 to −0.01||<0.001|
|Increased Cr, 12 months||−0.866||−1.31 to −0.42||<0.001|
|Increased CrCl, 12 months|| 0.030|| 0.02 to 0.04||<0.001|
Multivariate analysis for 12 month Hb is shown in Table 3. Older recipient age, female gender, older donor age, chronic infection and decreased renal function at 12 months all remained independently associated with lower Hb levels. In addition, the use of sirolimus remained a significant correlate with lower Hb at 12 months (slope =−1.06, 95% CI: −1.76 to −0.36, p = 0.003). Substituting serum creatinine for estimated glomerular filtration rate gave similar results.
Table 3. Linear regression analysis for Hb at 12 months
|Variable||Slope||95% CI of slope||p Value|
|Older age||−0.02||−0.04 to −0.01|| 0.018|
|Female gender||−1.40||−1.88 to −0.90||<0.001|
|African American||−0.28||−0.89 to 0.34|| NS|
|Kidney–pancreas|| 0.55||−0.23 to 1.33|| NS|
|Older donor age||−0.02||−0.04 to −0.01|| 0.023|
|Steroid wean|| 0.66||−0.08 to 1.40|| NS (0.08)|
|Sirolimus||−1.06||−1.76 to −0.36|| 0.003|
|Rejection||−0.52||−1.26 to 0.22|| NS|
|Infection||−1.02||−1.62 to −0.41|| 0.001|
|Increased CrCl, 12 months|| 0.01|| 0.01 to 0.02|| 0.02 |
|Angiotensin blockade|| 0.63|| 0.05 to 1.21|| 0.034|
On initial linear regression analysis the use of angiotensin blocking agents in the first year post-transplant was shown to be associated with higher Hb level at 12 months (Table 3). Forty-six patients (22% overall) were treated with ACEIs or ARBs during this time period. Eighteen (39%) of these patients had PTE and were started on angiotensin blockade after diagnosis. When PTE patients were excluded from analysis, the use of angiotensin blocking agents no longer had any significant association with Hb. Sirolimus use remained a significant independent predictor of lower Hb at 12 months by multivariate linear regression after excluding patients with PTE (slope =−0.67, 95% CI: −1.32 to −0.02, p = 0.041).
Table 4 shows the variables associated with PTE by univariate analysis. Older age, female gender, decreased creatinine clearance and use of sirolimus rather than MMF were all associated with a decreased incidence of PTE. Ethnicity, kidney versus kidney–pancreas transplant, rejection, infection, pre-existing disease, living versus cadaveric donor, type of calcineurin inhibitor and withdrawal of steroids were not associated with PTE. Logistic regression analysis is shown in Table 5. Older recipient age and female gender remained independently associated with a lower risk of PTE. Use of sirolimus rather than MMF also remained a negative correlate for PTE (odds ratio: 0.33, 95% CI: 0.12 to 0.89, p = 0.028).
Table 4. Variables associated with incidence of PTE by univariate analysis
|Variable||PTE||No PTE||p Value|
|Age||40 ± 10||47 ± 13||0.006|
| Female|| 8%||92%|| |
|CrCl, 12 mos||79 ± 28||69 ± 24||0.039|
Table 5. Logistic regression analysis for PTE
| Variable||Beta coefficient||Odds ratio (95% CI)|| p Value|
|Increased age (year)||−0.053||0.95 (0.91–0.98)||0.005|
|Female gender||−1.028||0.36 (0.14–0.92)||0.033|
|CrCl, 12 mos|| 0.004||1.00 (.99–1.0)||NS|
Of the 103 anemic patients at 6 months, 27 (26%) had iron studies, and an additional 20 (19%) were placed empirically on iron. Of patients tested, the rate of laboratory-confirmed iron deficiency was approximately 45% in both the MMF and sirolimus groups. Forty-one patients (19%) were treated with iron. Of these, 16 were on MMF and 25 were on sirolimus (13% vs. 29%, p = 0.003). The average time to treatment for both groups was 4 months. In addition, 18 patients were treated with erythropoietin (8% overall). Of these, 5 were on MMF and 13 on sirolimus (4% vs. 15%, p = 0.006). The average time to treatment with erythropoietin in both groups was approximately 5 months. Thirteen of 18 patients on erythropoietin were also treated with iron. Dosage of erythropoietin averaged 13 000 units per week in the MMF group and 18 000 units per week in the sirolimus group (p = NS).
We measured the response to treatment by evaluating the change in Hb from 6–12 months in patients who were treated with iron, erythropoietin or both (Figure 2). In the MMF group (n = 18), Hb significantly increased from 11.2 ± 1.1 to 12.0 ± 1.4 g/dL (p = 0.008). In the sirolimus group (n = 28), mean Hb remained unchanged (11.1 ± 2.1 at 6 months and 11.1 ± 2 g/dL at 12 months). Similarly, the incidence of anemia at 6 and 12 months decreased in the MMF group from 78% to 56% (p = 0.04), and remained unchanged in the sirolimus group (71–82%, p = NS).
Figure 2. Six and 12 month Hb levels (g/dL) for patients on sirolimus (n = 28) and MMF (n = 18) who were treated with iron and/or erythropoietin.
Download figure to PowerPoint
- Top of page
In this single center analysis, we sought to compare the effects of sirolimus- versus MMF-based immunosuppression on erythropoiesis following kidney transplantation. Our results suggest that anemia is more common, more severe and more resistant to treatment in patients receiving sirolimus than in those receiving MMF. The association of sirolimus therapy with lower Hb was independent of other factors known to impact negatively on erythropoiesis including older age, female gender, infection and decreased renal function. Our results show that treatment with sirolimus is also associated with a lower incidence of PTE.
Previous data on the influence of sirolimus on erythropoiesis has been limited. A recent multicenter study of kidney transplant recipients showed that anemia was associated with decreased renal function, donor age, angiotensin blocking agents, infection, lack of polycystic disease and treatment with MMF or azathioprine (19). Less than 2% of patients in that survey were treated with sirolimus. Kahan et al. reviewed 10 years of experience with sirolimus therapy, and described a dose-dependent association with anemia in phase I and II trials of the drug (22). Conversion from a calcinuerin inhibitor to sirolimus led to anemia in 72% of patients in a small European trial (23). Most of these patients were also treated with MMF or azathioprine.
Sirolimus may inhibit erythropoiesis at the level of the erythropoietin receptor (24). Binding of erythropoietin to its cytoplasmic receptor leads to the activation of a cascade of phosphorylating enzymes, including phosphoinositide 3-kinase (PI 3-kinase). PI 3-kinase is responsible for controlling cell survival and cell cycle progression in multiple cell lines, including erythroid precursors (25). One enzyme downstream form of PI 3-kinase is p70S6-kinase. This S6 kinase plays a role in mRNA translation in the cell and is regulated by mTOR. Through mTOR inhibition, sirolimus has been shown to block S6 kinase activity and consequently impair cell replication in an erythroid cell line (26).
This mTOR pathway may be particularly important for replenishing red cell precursors in a depleted state. Animal studies have shown minimal inhibition of erythrocytosis with sirolimus at steady state. However, in mice depleted of bone marrow cells by 5-fluorourocil, there was a significant impairment of hematopoietic recovery with sirolimus (27). Transplant patients are often recovering from anemia post-operatively, and a significant rise in endogenous erythropoietin has been shown early after transplantation (28). Our results suggest that sirolimus may create an ‘erythropoietin resistant’ state in some patients. The average erythropoietin dose in 13 sirolimus patients was 18 000 U/wk, or 210 U/kg/wk. Despite this high dose, patients did not respond with a rise in Hb. Further studies are needed to show if higher endogenous levels or exogenous dosages of erythropoietin are able to overcome anemia related to sirolimus.
The influence of sirolimus on the incidence of PTE may also be explained by these mechanisms. PTE is generally characterized by increased erythropoietin levels relative to hemoglobin, and by increased numbers of early erythroid precursors in the peripheral blood (29). If sirolimus creates a state of erythropoietin resistance, it may decrease the response of erythroid precursors to higher levels of circulating erythropoietin. In addition, insulin-like growth factor-1 (IGF-1) is elevated in patients with PTE and promotes erythropoiesis (30). The same mTOR-dependent S6 kinase is activated downstream from the IGF-1 receptor. Sirolimus has been shown to inhibit IGF-1 via S6 kinase inhibition (31).
A limitation of this study is its non-randomized retrospective design. There were obvious differences in the two patient populations in terms of gender, ethnicity and renal function. We attempted to control for these differences, but there are certain pre-transplant, post-transplant and donor variables for which we may not have accounted. Future prospective trials with randomization and intervention are needed to further define risk factors for anemia, including immunosuppressive medications such as sirolimus. Future studies are also needed to determine whether conversion from sirolimus to MMF or other immunosuppressants can be associated with improved erythropoiesis.
We also did not have consistent measurements of secondary causes of anemia such as iron studies. Iron deficiency is a common cause of anemia in the first year post-transplant, and underutilization of iron likely contributes to persistent anemia (32). We saw a significant improvement in anemia in MMF patients who were placed on iron, and it is likely that more patients who were anemic at 6 months would have benefited from supplementation. We did not see the same improvement in sirolimus patients who were treated with iron. It may be that they remained anemic despite repletion of iron stores, although further studies are required to confirm this. Finally, our retrospective analysis suggests a relative underutilization of erythropoietin in our anemic patients during the period of study.
In conclusion, we found significantly lower Hb levels with sirolimus compared to MMF, as well as a decreased incidence of PTE. The overall incidence of anemia was high and did not improve in sirolimus patients between 6 and 12 months. Anemia has been increasingly recognized as a problem post-transplant. Trials of prevention and treatment are needed to assess the impact of intervention and to determine if correcting anemia will have beneficial effects in terms of cardiovascular and overall morbidity and mortality.