- Top of page
Post-transplant anemia (PTA), a frequent complication during the first 3–6 months after transplant, is thought to be uncommon during the late post-transplant period. A study population of adults (> 18 years) transplanted during 1995 at Stanford University (n = 88) and University of North Carolina (n = 40) was selected. Data-collection points were 0, 1, 2, 3, 4 and 5 years post transplant. Anemia was defined as a hematocrit < 33 volume percentage. Thirty percent of patients were anemic at some time during the post-transplant period. The prevalence of PTA increased over time; by 5 years post transplant, 26% of the patients were anemic. Anemia occurred in 62.5% of patients converted from azathioprine to mycophenolate mofetil. A multivariate logistic regression model demonstrated a correlation between anemia and serum total CO2 (p = 0.002), BUN (p = 0.04), and creatinine (p = 0.045) at 1 year post transplant. At 5 years post transplant, only serum total CO2 (p = 0.0004) correlated with anemia. Thus, diminished renal excretory function and metabolic acidosis appear to be the most important correlates of late PTA. These findings should be interpreted in view of the fact that the newer immunosuppressive agents may have an even more profound effect on anemia and its recovery after transplantation.
- Top of page
Previous studies have reported that nearly all renal transplant recipients develop anemia during the early post-transplant period and experience resolution of the anemia by 3–6 months post transplant (1–6). Early PTA has been attributed to blood loss at the time of surgery, frequent blood draws, iron depletion (2, 4, 6–8), the persistent effect of uremic toxins (6), low erythropoietin (EPO) levels (6, 9, 10), EPO resistance (6, 9–12) and the negative effects of immunosuppression on erythropoiesis (13, 14).
The PTA that occurs late after transplantation has been attributed to decreased red blood cell survival (15), erythrocyte hypoproduction due to iron deficiency (2, 16, 17), low excretory graft function (16), and immunosuppressive agents (18–23). To date, there have been few studies of late PTA. In a cross-sectional study assessing the prevalence of late PTA, Saito et al. (24) reported a 23.3% prevalence of anemia (defined as Hgb: males, < 12.8 g/dL; females, < 11.5 g/dL). In the only longitudinal study (0–24 months post transplantation), in which mean hemoglobin levels were quantified relative to specific immunosuppressive regimens, Kahng et al. reported that anemia was corrected in most patients by 8 months post transplant and that early iron therapy, particularly in females, reduced the incidence of anemia (25). There were no differences between the mean hemoglobin levels in patients receiving dual immunosuppression with cyclosporine (CsA) + prednisone or triple immunosuppression with CsA + azathioprine (AZA) + prednisone. Furthermore, there was no significant downward trend in mean hemoglobin values over time. However, differing definitions of anemia and exclusion criteria can profoundly influence the reported prevalence of anemia.
We sought to expand on the previous reports by conducting a retrospective longitudinal analysis of adult renal transplant recipients, so that the prevalence, severity and predictors of PTA could be determined. We hypothesized that the prevalence of PTA would increase with time post transplant and in relationship to declining renal excretory function. The results of our data collection and analysis form the basis of this study.
Materials and Methods
A database was retrospectively created by chart review of adult (> 18 years of age) renal transplant patients who received transplants during the 1995 calendar year at Stanford University and at the University of North Carolina. 1995 was chosen because all of the patients had the potential of at least 5 years of follow-up at the time of analysis (March 2001). At Stanford University, 89 kidney transplants were performed during 1995. Twelve patients died and six patients lost their renal allografts prior to reaching 5 years post transplant. A total of 88 charts (99%) were retrieved for analysis. There were 44 adult renal transplants performed at the University of North Carolina in 1995. Seven adults had graft loses and another three died before 5 years. A total of 40 charts (91%) were available for review. Patients receiving combined organ transplants during 1995, including heart–kidney, liver–kidney and kidney–pancreas transplant recipients were excluded from analysis.
At any given timepoint, the patient's hematocrit (HCT) was assessed relative to the Dialysis Outcomes Quality Initiative (DOQI) HCT target guidelines for EPO therapy (26). According to DOQI guidelines, the target range for HCT is 33–36 volume per cent. We reasoned that a patient with a HCT < 33 volume per cent would be sufficiently anemic to be a candidate for diagnostic studies, iron and/or EPO therapy. For the purpose of this study, anemia was defined as a HCT < 33 volume per cent or as an EPO dependency to maintain a normal HCT. Recurrent anemia was defined as anemia at two or more yearly time-points during the post-transplant period.
Hematocrit values were also compared to age- and gender-specific normative values by calculating a HCT Z-score. The age-specific definitions of anemia used for this study are shown in Table 1 (27). The HCT Z-score was calculated using the following equation:
Table 1. : Normative age-adjusted HCT values
| ||Mean HCT||− 2 SD below the mean|
| 14–50 years||42||36|
| 50–80 years||40||36|
| 18–50 years||47||42.3|
| 50–60 years||45||40.5|
| 60–70 years||43||38.7|
| 70–80 years||40||36|
A diagnosis of anemia and HCT Z-scores were the primary variables analyzed. Other variables evaluated included the transplant center, age, gender, weight, donor type, serum total CO2, BUN, creatinine, calculated creatinine clearance (Ccr) and immunosuppressant medications. Data regarding antihypertensive, sodium bicarbonate, iron and erythropoietin (EPO) therapy were also entered. The laboratory results at the time of transplantation and closest to the specified post-transplant interval of 1, 2, 3, 4 and 5 years post transplant were selected for analysis. The Ccr was calculated using the Cockcroft-Gault equation (28).
Continuous variables were compared using the unpaired Student t-test, and nominal data were compared using Fisher analysis. manova testing was used to analyze the relationship between continuous and categorical variables Patients were sorted into anemic and non-anemic based on the parameters defined above. Univariate analysis of 1- and 5-year variables was performed using Pearson correlation coefficients. A multivariate logistic regression model using 1- and 5-year post-transplant data was created to control for the effects of other predictors. Variables used in the model were those with p-values < 0.05 identified in the univariate analysis. Using a backward elimination procedure, the least significant variables were dropped from the model stepwise until all variables remaining in the model had a p < 0.05. In all cases, the significance level was set at p ≤ 0.05.
- Top of page
Baseline patient characteristics at the time of transplantation are shown in Table 2. There were significant differences between the demographics of the transplant recipients at Stanford University and at the University of North Carolina. There were differences in ethnic composition of the study populations between the two centers (p < 0.0001) The Stanford University group had a greater percentage of Asian (p = 0.0004) and Hispanic patients (p < 0.0001), while African Americans constituted a greater percentage of renal transplant recipients at the University of North Carolina (p < 0.0001). The percentage of patients who were anemic at baseline was nearly identical at the two centers, 40% at the University of North Carolina and 44% at Stanford University. Eighty-five percent of the University of North Carolina patients and 84% of the Stanford University patients had HCT Z-score of > 2 standard deviations below the mean. The initial immunosuppression therapy was very different at the two institutions. Both institutions predominantly used cyclosporine (CsA), prednisone and azathioprine. At the time of transplant, Stanford University transplant recipients were more likely to receive a combination of CsA, prednisone and mycophenolate mofetil (MMF) (p = 0.002). Patients at the University of North Carolina were more likely to receive dual therapy of CsA and prednisone (p < 0.0001) or tacrolimus and prednisone (p = 0.002). There were small numbers of patients at each institution who received tacrolimus-based immunosuppression with MMF with or without prednisone (Table 2). Delayed graft function was more common at Stanford University (p = 0.006).
Table 2. : Patient characteristics at transplantation
| ||Combined||Stanford University||University of North Carolina||p-value|
|Age|| 44.4 ± 11.8||44.9 ± 11.8||43.3 ± 12.0||0.47|
|Gender (F:M)|| 57/71 (45%)||42/46 (48%)||15/25 (38%)||0.21|
| African American|| 24 (19%)|| 8 (9%)||16 (40%)||< 0.0001|
| Asian|| 24 (19%)||23 (26%)|| 1 (3%)||0.001|
| Caucasian|| 64 (50%)||42 (48%)||22 (55%)||0.34|
| Hispanic|| 12 (9%)||12 (14%)|| 0 (0%)||< 0.0001|
| Other|| 4 (3%)|| 3 (3%)|| 1 (3%)||0.85|
|DGF|| 19/109||18/70|| 1/39||0.006|
|HCT|| 33.3 ± 5.1||33.4 ± 5.3||33.0 ± 4.7||0.65|
|HCT Z-score||− 4.38 ± 2.21||− 4.20 ± 2.12||− 4.74 ± 2.36||0.22|
| CsA + Pred + AZA|| 68 (53%)||45 (51%)||23 (58%)||0.41|
| CsA + Pred + MMF|| 24 (19%)||23 (26%)|| 2 (5%)||0.002|
| CsA + Pred|| 15 (12%)|| 3 (3%)||10 (25%)||< 0.0001|
| Tacrolimus + Pred + MMF|| 2 (2%)|| 2 (2%)|| 0 (0%)||< 0.0001|
| Tacrolimus + Pred|| 7 (5%)|| 2 (2%)|| 5 (13%)||< 0.0001|
| Tacrolimus + MMF|| 1 (1%)|| 2 (2%)|| 0 (0%)||< 0.0001|
| Tacrolimus + Pred + AZA|| 2 (2%)|| 2 (2%)|| 0 (0%)||< 0.0001|
| Unknown|| 9 (7%)|| 9 (10%)|| 0 (0%)||< 0.0001|
The mean duration of follow-up for the combined study group was 3.7 ± 1.9 years. Anemia was more prevalent (41%) at the time of transplant than at any other time point (Table 3). Thirty-eight of the 128 patients (30%) were anemic at some time during the post-transplant period. Nineteen of the 38 anemic patients (50%) experienced recurrent anemia. There was a lower prevalence of anemia during the first 2 years post transplant, but the prevalence and severity of anemia increased over time, so that 26% of patients were anemic, and 53% of patients had low HCT Z-scores (more than 2 standard deviation below the mean) at 5 years post transplant. The severity of the anemia (HCT Z-scores) relative to time post transplant is shown in Figure 1.
Table 3. : Prevalence of anemia and mean HCT Z-score
|Years post transplant||0||1||2||3||4||5|
|Mean HCT Z-score||− 4.38 ± 2.21||− 1.29 ± 2.57||− 1.35 ± 2.56||− 1.85 ± 2.16||− 2.16 ± 2.48||− 2.42 ± 2.73|
| n =|| 51|| 12|| 12||12||15||19|
| Percent|| 43|| 12|| 12||14||18||26|
|Low HCT Z-score*|
| n =||101|| 36|| 43||35||38||39|
| Percent|| 84|| 35|| 43||40||46||53|
|Total (n =)||120||103||100||87||83||73|
Figure 1. Severity of anemia relative to the time of transplantation. Hematocrit Z-scores are plotted relative to years post transplantation. A light gray line demarcates those patients who were anemic, > 2 standard deviations below the mean.
Download figure to PowerPoint
The 1- and 5-year post-transplant variables were compared for anemic and non-anemic patients (Table 4). Anemic patients had lower serum total CO2 values (p = 0.019) and lower mean Ccr values (p = 0.005), at 1 year post transplant. While delayed graft function (DGF) did not correlate with diminished Ccr (p = 0.27), DGF correlated with anemia (p = 0.01). Anemic patients were more likely to have Ccr values < 50 mL/min/1.73 m2 than non-anemic patients. At 5 years post transplant, anemic patients had higher BUN values (p = 0.005) and lower serum total CO2 (p < 0.0001) and calculated creatinine clearance (CCr) values (p = 0.033) and were more likely to be receiving MMF therapy (p = 0.003). As with the 1-year post-transplant data, anemic patients were more likely to have Ccr values < 50 mL/min/1.73 m2; 51% of anemic patients compared to only 15% of the non-anemic patients. Yet, poor renal excretory function was not an absolute indicator of anemia. There were anemic patients at 1 (n = 1) and 5 (n = 2) years post transplant who had Ccr values > 75 mL/min/1.73 m2. Conversely, there were patients at both time-points who did not have anemia, but had poor renal excretory function with Ccr < 50 mL/min/1.73 m2.
Table 4. : Comparison of anemic and non-anemic patients at 1 year and 5 years post transplant
| ||1 year post transplant||5 years post transplant|
| ||Anemic (n =12)||Non-anemic (n =91)||p-value||Anemic (n =19)||Non-anemic (n =54)||p-value|
|Age||45.1 ± 12.2||44.6 ± 23.2||0.89||45.7 ± 10.8||51.0 ± 10.8||0.07|
|Gender (F:M)|| 7/5||43/48||0.45|| 8/11||25/29||0.71|
| African American|| 2||15||1|| 8|| 6||0.009|
| Asian|| 3||17||0.55|| 2||12||0.58|
| Caucasian|| 4||50||0.13|| 8||31||0.20|
| Hispanic|| 2|| 7||0.037|| 1|| 4||0.50|
| Other|| 1|| 2||–|| 0|| 1||–|
|LD/CAD|| 2/10||32/59||0.18|| 6/13||17/37||0.39|
|UNC/Stanford|| 2/10||30/61||0.22|| 6/13||20/34||0.72|
|Total CO2||20.1 ± 4.6||24.0 ± 3.2||0.019||19.3 ± 3.0||25.6 ± 2.7||< 0.0001|
|BUN||31.3 ± 9.5||27.0 ± 16.8||0.20||59.1 ± 40.8||29.1 ± 16.5||0.005|
|Creatinine|| 2.9 ± 2.6|| 1.6 ± 0.7||0.12|| 4.5 ± 5.1|| 1.8 ± 1.5||0.033|
|Mean||43.9 ± 23.8||68.5 ± 25.7||0.005||43.1 ± 29.6||65.8 ± 28.5||0.008|
|> 75 mL/min/1.73 m2|| 1(8%)||29(32%)||0.082|| 2(20%)||14(32%)||0.07|
|50–75 mL/min/1.73 m2|| 4(33%)||38(42%)||0.558|| 4(29%)||23(53%)||0.002|
|< 50 mL/min/1.73 m2|| 7(58%)||23(26%)||0.004||12(51%)||17(15%)||< 0.0001|
| Tacrolimus/CsA|| 4/8||11/79||0.13|| 7/11||13/40||0.50|
| MMF/AZA|| 3/2||19/44||< 0.0001|| 9/3|| 8/23||0.003|
Gender, ethnicity, donor type, hypertension and immunosuppression therapy did not influence HCT Z-scores at 1 year post transplant. However, at 5 years post transplant, males were more likely to have lower HCT Z-scores than females: − 3.27 ± 2.72 vs. −1.39 ± 2.40 (p = 0.0028). African Americans had lower HCT Z-scores when compared to Asians (p = 0.017) and Caucasians (p = 0.02). There was no statistically significant difference in the HCT Z-scores of CsA- vs. tacrolimus-based immunosuppression groups at 1 or 5 years post transplant.
Because renal excretory function was thought to influence the prevalence of PTA anemia, Ccr was evaluated for the combined study group and special groups of interest. For the combined group, Ccr was fairly stable over the first 4 years, 65.6 ± 26.6 mL/min/1.73 m2 at 1 year, 66.9 ± 26.1 mL/min/1.73 m2 at 2 years, 67.0 ± 26.5 mL/min/1.73 m2 at 3 years, 65.2 ± 29.2 mL/min/1.73 m2 at 4 years post transplant, but decreased to 60.1 ± 30.2 mL/min/1.73 m2 5 years post transplant. At 1 year post transplant, there was no correlation between Ccr and either gender, age (< 55 years vs. > 55 years of age), donor type, or CsA- or tacrolimus-based immunosuppression regimens. However, Ccr was greater in African Americans transplant recipients than Asians (76.5 ± 35.1 compared to 56.1 ± 18.9 mL/min/1.73 m2, p = 0.02). At 5 years post transplant, there was no correlation between Ccr and gender, donor type, ethnicity, age, or CsA- or tacrolimus-based immunosuppression regimens.
To determine whether MMF treatment correlated with anemia, intention to treat analysis of CsA + Prednisone + MMF vs. CsA + Prednisone + AZA immunosuppression was performed. With the exception of a higher prevalence of anemia in the MMF-treated group 4 years post transplant (35.7 vs. 10.2%, p = 0.04), there was no difference in mean HCT Z-scores or diagnosis of anemia between groups at any time-point post transplant. We also analyzed whether conversion to MMF correlated with anemia. Of the eight patients converted from AZA to MMF, five (62.5%) met the DOQI criteria for anemia.
Iron therapy was infrequently used in the management of PTA. Of the 12 anemic patients at 1 year post transplant, iron therapy was prescribed for seven (58%). Eleven of the 91 (12%) non-anemic patients were receiving iron therapy. Only one (8%) of the anemic patients was receiving EPO therapy. At 5 years post transplant, none of the 19 anemic patients was receiving EPO therapy and only two (11%) patients were receiving iron therapy. Six per cent of the non-anemic patients were receiving iron therapy.
Univariate analysis was performed comparing HCT Z-scores to all variables. At 1 and 5 years post transplant, HCT Z-scores correlated with serum total CO2 (Year 1: 0.391, p < 0.0001 and Year 5: 0.645, p < 0.0001), BUN (− 0.200, p = 0.044 and − 0.547, p < 0.0001), creatinine (− 0.308, p = 0.0015 and − 0.400, p < 0.0004) and Ccr (0.234, p = 0.0175 and 0.371, p = 0.0012).
Multivariate logistic regression was performed using the significant continuous variables identified by univariate regression and all nominal variables. Results of the multivariate logistic regression for 1 and 5 years post transplant are shown in Tables 5a and 5b, respectively. Serum total CO2, BUN and creatinine were identified as the variables which correlated best with anemia at 1 year post transplant. At 5 years post transplant, serum total CO2 correlated best with the diagnosis of anemia (Figure 2).
Table 5a. : Multivariate logistic regression analysis at 1 year post transplant
| ||Coefficient (± 1 SEM)||p-value|
|Intercept||6.475 ± 3.213||0.04|
|Total CO2||− 0.402 ± 0.131||0.0021|
|BUN||0.069 ± 0.034||0.0392|
|Creatinine||1.136 ± 0.567||0.0451|
Table 5b. : Multivariate logistic regression analysis at 5 years post transplant
| ||Coefficient (± 1 SEM)||p-value|
|Intercept||16.813 ± 4.934||0.0007|
|Total CO2||− 0.795 ± 0.225||0.0004|
Figure 2. Bivariate graph of serum total CO2 vs. HCT Z-score in patients 5 years post transplant. Hematocrit Z-scores are plotted relative to serum total CO2.
Download figure to PowerPoint
- Top of page
Post-transplant anemia was a common problem which affected 30% of the study population at least once during the first 5 years after transplant. Recurrence of PTA was common (50%) and the prevalence and severity of anemia increased with time after transplantation. The prevalence of PTA in our study was similar to that reported by Saito et al. (24), who reported a 23.3% prevalence of anemia in a cross-sectional analysis of renal transplant recipients with a creatinine < 3 mg/dL.
The choice of clinically relevant age-appropriate normative HCT values can profoundly alter study results. Most current definitions of anemia reflect population-based statistical analysis, so that an otherwise healthy individual is defined as anemic when her/his HCT is > 2 standard deviations below the mean for age and gender. There are differences in the published normative values for adults because of differences in populations and analysis techniques (27, 29–32). There is currently no workable definition of anemia which is based on HCT-induced changes in cardiac and pulmonary physiologic responses, cognition or exercise tolerance. In this study, HCT Z-scores were calculated at every time-point for each study patient. When hematocrits were compared to normative values, many men had hematocrits > 2 standard deviations below the mean. For example, a HCT value of 42.0 volume per cent is > 2 standard deviations below the mean for males 18–50 years of age; yet, such a value would rarely be deemed clinically significant by a transplant physician. Because of our desire to select a definition of anemia which was clinically significant, DOQI guidelines were chosen. The DOQI guidelines have established a target HCT range of 33–36 volume per cent for adult dialysis patients receiving EPO therapy, based on medical evidence which shows improved clinical outcomes (26). Nevertheless, it seems prudent for transplant physicians to be aware that their patients may experience the adverse physiologic effects of anemia despite having a HCT within or above the DOQI target range.
As expected, renal excretory function correlated with anemia in adult renal transplant recipients. Yet, the effect of renal excretory function on anemia was not absolute, as a small number of anemic patients had Ccr values > 75 mL/min/1.73 m2, while some non-anemic patients had CCr < 50 mL/min 1.73 m2. A previous study by van Dullemen et al. demonstrated that EPO production decreases and EPO resistance increases relative to the decline in renal excretory function (33). Even without a significant decline in renal excretory function, DGF, acute tubulointerstitial rejection (7, 34), chronic rejection (34) and possibly long-term calcineurin inhibitor toxicity (35) may cause diminished EPO production.
Since renal tubular injury attributable to DGF, allograft rejection and calcineurin inhibitor toxicity is common in renal transplant recipients, it seemed reasonable to evaluate serum total CO2 as an important indicator of renal tubular function. Indeed, serum total CO2 was an important correlate with anemia, both 1 and 5 years post transplantation. We speculate that the presence of metabolic acidosis is a surrogate indicator of renal tubular injury, diminished renal excretory function and diminished EPO production.
It was surprising that few anemic patients were ever treated with EPO or iron therapy. This finding may demonstrate a need for increased awareness of anemia and increased concern about subnormal HCTs by the patient's primary nephrologists and transplant physicians. Treatment of PTA appears to be important, as EPO-mediated correction of the hemoglobin, from 7.89 ± 1.04–10.26 ± 1.84 g/dL, has been shown to cause a demonstrable improvement in the quality of life, as assessed by use of the General Sickness Impact Profile and the disease-specific Transplant Disease Questionnaire in adult renal transplant recipients (36).
Determining the relative contribution of any specific or combined immunosuppression therapy to the development of anemia is fraught with problems when a retrospective database is used. When changes in immunosuppression combinations occur, determining the relationship between the immunosuppression change and subsequent anemia is difficult. Immunosuppression drugs are also used in combination, so determining the relative contribution of a specific medication is problematic. Mycophenolate mofetil therapy at 5 years post transplant correlated with anemia in the univariate analysis. Mycophenolate mofetil has been previously reported as a cause of post-transplant anemia (13). However, the correlation of anemia and MMF cannot be attributed to patients that were started on MMF therapy at the time of transplant, but due to the conversion to MMF therapy in individuals who had acute or chronic rejection episodes. Given the rather small patient population in our study, further evaluation of the effect of MMF on the HCT relative to other immunosuppressive medications seems warranted.
In conclusion, PTA in adult renal transplant recipients is a common complication that is under-recognized. Low serum total CO2 values, diminished renal excretory function appear to be important predictors. Immunosuppression in general, and MMF specifically, may play a role in late PTA, but further analyses with larger patient populations are required. Physicians should monitor their transplant patients for anemia, and initiate appropriate diagnostic studies and consider iron and/or EPO therapy if indicated. Because of the relatively small number of patients in our study (n = 128), further study is needed to more completely determine the causative factors and health implications of post-transplant anemia.