Background: The association of anemia with outcomes after renal transplantation (RT) is unclear.
Methods: We performed a retrospective study that included patients who received a RT in Spain in 2007. We collected data on anemia (hemoglobin [Hb] <11 g/dL and/or erythropoietic agents and/or transfusion in the previous month) as well as transplantation and clinical data during follow-up. We used multivariate Cox models to predict graft and patient survival.
Results: We included 639 patients; 7.2% lost their graft and 6.3% died. The prevalence of anemia was 84% at 7 days, 77% at 1 month, 41% at 2 months, 16% at 12 months, 14% at 24 months, and 18% at 36 months. After adjusting by glomerular filtration rate (hazard ratio [95% confidence interval], 0.96 [0.93–0.98]), low Hb levels at 1 month remained as an independent predictor of graft loss (hazard ratio for each 1 g/dL increase, 0.72 [0.54–0.96]) along with a maximum panel-reactive antibody of more than 10% (3.80 [1.73–8.36]), a donor with stroke (3.30 [1.31–8.28]), and one or more acute rejection episode (13.89 [4.78–40.37]). Tacrolimus use was a protective factor (0.24 [0.11–0.50]).
Conclusions: Low Hb levels in the early posttransplantation period (1 month) seem to be an independent prognostic factor for graft loss, but not for mortality, in Spanish RT patients regardless of graft function, recipient and donor characteristics, unfavorable events within the first month, and immunosuppression.
1 Hospital del Mar, Barcelona, Spain.
2 Hospital Universitario La Paz, Madrid, Spain.
3 Hospital Universitario General de Alicante, Alicante, Spain.
4 Hospital Universitario Virgen de la Arrixaca, Murcia, Spain.
5 Hospital Son Espases, Mallorca, Spain.
6 Hospital Universitario Virgen del Rocío, Sevilla, Spain.
7 Complejo Hospitalario Universitario Santiago, Santiago de Compostela, Spain.
8 Ciutat Sanitaria i Universitaria de Bellvitge, l’Hospitalet de Llobregat, Spain.
9 Hospital Regional Universitario Carlos Haya, Málaga, Spain.
10 Hospital 12 de Octubre, Madrid, Spain.
11 Hospital Universitario Dr. Peset, Valencia, Spain.
12 Hospital General Vall d’Hebron, Barcelona, Spain.
13 Hospital Universitario Marques de Valdecilla, Santander, Spain.
14 Hospital Clínic, Universidad de Barcelona, Institut d’Investigacions August Pi i Sunyer (IDIBAPS), Barcelona, Spain.
The study was conducted under the auspice of the Spanish Society of Transplantation. This study was supported in part by a grant from Amgen.
Amgen did not have any role in the study design; collection, analysis, and interpretation of data; writing of the article; and the decision to submit the article for publication. The authors declare no other conflicts of interest.
Address correspondence to: Dr. Julio Pascual, Nephrology Service, Hospital del Mar, Passeig Marítim de la Barceloneta, 25, ES-08003 Barcelona, Spain.
J.P., J.M.G., D.H., J.M.M., L.M.M., D.S., M.A., and J.M.C. designed the study. M.A. and J.M.C. coordinated the working group. J.P. and D.H. reviewed the analysis. All authors contributed to clinical data collection, reviewed the article, and approved its final version.
Received 14 February 2013. Revision requested 11 March 2013.
Accepted 4 June 2013.
This is an open-access article distributed under the terms of the CreativeCommons Attribution-NonCommercial-NoDerivatives 3.0 License, where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially. http://creativecommons.org/licenses/by-nc-nd/3.0.
Accepted August 1, 2013
Anemia is a common finding during the first months after renal transplantation (RT), with a prevalence of approximately 40% (1). Several cutoff points have been used to define posttransplantation anemia, but the limit of hemoglobin (Hb) level below 11 g/dL is the currently proposed target for anemia therapy according to the National Kidney Foundation Kidney Disease Outcomes Quality Initiative recommendations (2, 3) and constitutes a minimum threshold accounted for in all definitions. However, any evaluation of anemia should also include patients with Hb levels successfully corrected by erythropoiesis-stimulating agents (ESA) use or red blood cell (RBC) transfusion, who are also physiologically anemic.
The usual pattern of posttransplantation anemia is a decrease in Hb levels during the first months after RT (early-onset anemia), an increase approximately at 1 year, and then a second decline coinciding with the deterioration of graft function (late-onset anemia) (4, 5). Several factors may determine the development of early anemia: pretransplantation anemia, blood loss during surgery, and low erythropoietin production by the graft, which does not begin to synthesize erythropoietin until it reaches an adequate renal function (6, 7). Iron deficiency can also induce or aggravate preexisting anemia (8). Other proposed factors are the use of immunosuppressive agents such as azathioprine, mycophenolate mofetil, or sirolimus, the use of angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB), alterations in iron homeostasis, or the advanced age of recipient and donor (9).
Posttransplantation anemia has been associated with increased cardiovascular morbidity and mortality as well as a deterioration or loss of graft function. However, the evidence in this regard is contradictory, and there is no clear consensus on its management (5, 10).
The objectives of our study were to describe the prevalence of early-onset anemia after RT; to describe its relationship with recipient, donor, transplantation, and posttransplantation management variables; to describe the management and clinical evolution of posttransplantation anemia during the first 3 years after transplantation; and to explore the relationship between early-onset anemia and transplant outcomes (renal function, death-censored graft survival, and patient survival).
Among 675 patients fulfilling selection criteria, 36 were excluded due to lack of information regarding Hb levels and/or anemia treatment in the first month after transplantation. A total of 639 patients were analyzed, with a follow-up time of 1787 patient-years, and a mean (SD) follow-up time of 2.7 (0.7) years.
Anemia Presence and Management Over Time
After transplantation, the prevalence of anemia (Hb <11 g/dL and/or any ESA use or at least one RBC transfusion within the last 4 weeks) was maximum at 7 days and decreased slowly and progressively during the first 12 months, with a slight rebound at 36 months (70.3% pretransplantation, 84.0% at 7 days, 78.0% at 14 days, 77.2% at 1 month, 41.1% at 2 months, 34.0% at 3 months, 26.3% at 6 months, 16.1% at 12 months, 14.1% at 24 months, and 18.5% at 36 months). The percentage of patients treated with ESAs was 62.1% pretransplantation, 52.6% at 7 days, 49.6% at 1 month, 28.7% at 3 months, 22.6% at 6 months, 13.8% at 12 months, 12.6% at 24 months, and 15.2% at 36 months. Overall, 56% of patients received iron supplementation and 42% received at least one RBC transfusion. Among patients with ESAs, the frequency of noncorrected Hb levels (<11 g/dL) was 14.4% at baseline, 75.2% at 7 days, 41.9% at 1 month, 18.6% at 3 months, 14.1% at 6 months, 17.2% at 12 months, 11.0% at 24 months, and 32.9% at 36 months.
The concordance between pretransplantation anemia and early-onset anemia was very high: only 11.5% of anemic patients at 1 month were nonanemic before transplantation. Anemia prevalence throughout the follow-up was significantly higher in patients who had developed early-onset anemia (data not shown). The concordance between early-onset anemia and late-onset anemia was also very high: 79%, 75%, and 85% of patients with anemia at 12, 24, and 36 months, respectively, were anemic at month 1. The concordance between anemia at 1 year and anemia at 2 years was moderate: 57% of patients with anemia at 1 year are also anemic at 2 years.
The characteristics and clinical outcomes of the included patients according to the presence of anemia at 1 month are displayed in Table 1. There were only 8 of 493 (1.6%) patients with anemia at 1 month who had normal glomerular filtration rate (GFR) levels (≥90 mL/min). Patients with anemia at 1 month had a significantly higher incidence of surgical complications and reinterventions, postsurgical bleeding, delayed graft function, acute rejection, serious infection, graft loss, and death.
Predictors of Early-Onset Anemia
Independent risk factors for early-onset anemia were renal function at month 1 (GFR; hazard ratio [HR], 0.97; 95% confidence interval [CI], 0.96–0.98; P<0.001) and several recipient-related (maximum panel-reactive antibody [PRA] >10%: HR [95% CI], 2.64 [1.28–5.44]; P=0.008) and immediate posttransplantation-related variables (postsurgical bleeding: HR [95% CI], 3.24 [1.34–7.83]; P=0.009; delayed graft function: HR [95% CI], 1.59 [1.03–2.46]; P=0.037).
Anemia and Renal Function Over Time
A significant correlation was observed between Hb levels and renal function after transplantation (GFR) at all time points (data not shown). Among all time points, the presence of anemia at 1 month was the most strongly associated to evolution of graft function over time (Fig. 1).
In the repeated-measures model, the association between anemia and renal function was not homogeneous across time. A significant interaction was found between time after transplantation and the composite variable “Hb levels and ESA treatment” (Table 2): within the first month, there was a negative association with GFR in untreated or nonresponding anemic patients. Between 6 and 12 months, there were no relevant differences in GFR levels according to Hb levels and ESA treatment. At 36 months, anemic patients treated with ESA (with Hb <11 or ≥11 g/dL) had worse renal function than nonanemic patients.
Early-Onset Anemia and Graft Loss
In the univariate analyses testing the influence of several variables related to early-onset anemia and graft loss, both anemia presence at 1 month (HR, 5.40; P=0.005) and 2 months (HR, 4.54; P<0.001) and continuous Hb levels (observed levels, independently from ESA treatment or RBC transfusion) at 1 month (HR for each 1 g/dL increase, 0.59; P<0.001) and 2 months (HR, 0.71; P=0.001) displayed a strong and significant association. Patients with Hb levels less than 11 g/dL at 1 month with (HR, 5.14; P=0.025) or without (HR, 6.17; P=0.004) ESA treatment (anemia nonresponders and untreated anemics, respectively) had an increased risk of graft loss compared with nonanemics. Significant associations were observed between patients with low serum iron levels at 2 months (HR for each 1 μg/dL increase, 0.96; P=0.003) or receiving iron supplementation at 1 month (HR, 2.48; P=0.015) and graft loss. No significant results were obtained for pretransplantation Hb levels, ESA use at 1 or 2 months; propensity score for post-RT ESA use; pretransplantation and posttransplantation (first month) ESA use; and ferritin and transferrin saturation index (TSI) levels at 2 months (data not shown).
In the multivariate analyses, among all these variables, only continuous Hb levels at 1 month remained as an independent predictor of graft loss. After adjusting by GFR at 1 month, lower Hb levels were associated with shorter graft survival along with a maximum PRA more than 10%, a donor with stroke, and the presence of multiple acute rejection. Tacrolimus use was a protective factor (Table 3).
Early-Onset Anemia and Death
Although a trend was observed in the univariate analysis (Table 1; Fig. 2), no significant association was found in the multivariate analysis between anemia at 1 month (HR [95% CI] for each 1 g/dL increase, 1.09 [0.82–1.37]; P=0.439) or any other related variable (continuous Hb levels, composite variables [Hb levels and ESA use at 1 month; pretransplantation and early posttransplantation ESA use]) and risk of death. The independent predictor factors of mortality were older recipient’s age (HR [95% CI] for each 1-year increase, 1.05 [1.01–1.09]; P=0.006), cardiovascular disease before transplantation (HR [95% CI], 3.37 [1.75–6.49]; P<0.001), serious infection after transplantation (HR [95% CI], 2.53 [1.24–5.15]; P=0.011), and low GFR levels at 1 month (HR [95% CI] for each 1 mL/min/1.73 m2 increase, 0.97 [0.95–0.99]; P=0.010).
Although early anemia is a frequent finding after RT, there are limited data on its prevalence, characteristics, and influence on graft and patient survival. Most currently available data come from cross-sectional studies that have examined the prevalence of anemia in heterogeneous populations of patients with different years of follow-up (11). The present study has analyzed a broadly representative cohort of patients, which allows to describe accurately the presence and management of anemia and to analyze its relationship with the clinical course over approximately 3 years after transplantation. According to the data from the Spanish National Transplant Organization (12), the total number of RTs in 2007 in Spain was 2139. Thus, our sample constitutes approximately 30% of all transplanted patients in that year.
In line with previous studies (4, 5), we observed a very high prevalence of anemia within the first month after transplantation (approximately 3 of 4 patients), which decreased quickly within the second month, and more slowly later on, achieving minimum values (approximately 15%) between 1 and 2 years. At 3 years, the prevalence of anemia increased again, coinciding with a slight deterioration of graft function in some patients. As expected, early-onset anemia was strongly associated with unfavorable events in the immediate posttransplantation period (postsurgical bleeding and delayed graft function). The high concordance between pretransplantation and early posttransplantation anemia and the association found with a maximum PRA higher than 10% suggest that some recipient-related factors such as sensitization could be involved in the development of posttransplantation anemia. However, we cannot discard that the association with a high PRA is due to a cumulative effect of several anemia-predisposing factors in these patients (use of polyclonal antibodies, high doses of mycophenolate mofetil [not collected], higher incidence of delayed graft function, and acute rejection). Like previous studies, we found no relationship between early-onset anemia and any immunosuppressive drug or with ACEI or ARB use (13). We observed an apparent relationship of anemia with low sideremias (without difference in TSI or ferritin), similar to previous studies (14). The albumin levels were also lower in anemic patients. The low serum iron and hypoalbuminemia may reflect more a situation of malnutrition and/or inflammatory state rather than iron deficiency.
Patients with early-onset anemia had a significantly poorer graft function over the entire follow-up and a higher incidence of acute rejection, serious infection, graft loss, and death. Due to the observational nature of the study, we cannot infer a causal relationship for these observations. However, we conducted an extensive adjustment for confounding variables in the multivariate analyses to investigate if the presence or management of early-onset anemia may have a long-term effect on graft and patient survival.
In the analysis of graft survival, independently of anemia treatment, low Hb levels at 1 month remained as an independent predictor of graft loss. This observation agrees with previous studies, although it must be noticed that all of them analyzed the effect of “late-onset” anemia (15–18). We noticed a strong correlation between early- and late-onset anemia, which suggests that some patients remain anemic most of the time, but obviously our analysis included a broader group of anemic patients than that considered by the aforementioned studies.
In the analysis of patient survival, although a trend was observed, it disappeared after adjusting by confounders, which make us discard any independent effect on mortality.
Currently, there is no consensus regarding the administration of ESAs in the early posttransplantation period. A recent publication in the official journal of the Spanish Society of Nephrology (19) recommends that, in patients with Hb less than 11 g/dL in two consecutive measurements separated by 15 days, iron supplementation should be started if the patient has ferritin levels less than 100 pg/mL and/or TSI less than 20% and/or percentage of hypochromic erythrocytes more than 10%. Once the iron deposits have been repleted, if the anemia persist, treatment with an ESA should be initiated at the following doses: darbepoetin-α, 0.45 to 0.60 μg/kg/week (adjust to 0.75 μg/kg every 2 weeks); epoetin-α/β, 50 IU/kg three times per week; and continuous erythropoietin receptor activator, 0.6 μg/kg every 2 weeks (adjust to 1.2 μg/kg/month).
Patients who initiated an immunosuppressive regimen including tacrolimus within the first week after transplantation had a significantly lower risk of graft loss. This association was independent from that of anemia, because the prevalence of anemia in patients with and without tacrolimus was the same. Patients who did not receive tacrolimus (<15%) had some unfavorable characteristics: a longer time on dialysis, older recipient and donor age, more stroke as cause of donor’s death, a higher percentage of diabetes or vascular disease as chronic kidney disease etiology, lower albumin levels, and a higher percentage of delayed graft function (data not shown). Although the association of tacrolimus with better graft outcomes could be due to a combination of all these factors, the fact that all of them were taken into account in the multivariate analysis suggests that a direct protective effect cannot be ruled out for this immunosuppressive agent (20).
Our study has some limitations. Because the Modification of Diet in Renal Disease equation overestimates GFR in patients with mild renal dysfunction (21), we cannot discard that the observed influence of Hb levels at 1 month on graft survival is not truly independent and that low Hb levels are an indirect marker of impaired graft function. We were unable to discount the possibility of confounding factors not accounted for in the adjustments made or of residual confounders due to measurement error. Because the study was observational, we could not establish a causal relationship between low Hb levels at 1 month and shorter graft survival, nor can we rule out the possibility that these levels are associated with other unmeasured factors that are causally related to poorer outcomes.
Strengths of our study include the large sample size, unbiased recruitment representing the Spanish RT population, the exhaustive data collection, and an extensive analysis that accounted for all collected variables in the multivariate models.
In conclusion, the prevalence of anemia after RT is very high within the first 2 months. It decreases progressively until approximately 15% between 1 and 2 years and shows a slight rebound at 3 years. Some peritransplantation factors, such as highly sensitized recipients, postsurgical bleeding, and delayed graft function, are associated with early-onset anemia, which may help to identify subjects who could beneficiate from ESA treatment. ESAs are administered to approximately 50% of patients in the first month after transplantation. Our results suggest that low Hb levels in the early posttransplantation period (1 month) could be an independent prognostic factor for graft loss, but not for mortality, in Spanish RT patients regardless of early graft function, recipient and donor characteristics, unfavorable events within the first month, and immunosuppression. Patients who have their anemia corrected with ESAs in the early posttransplantation period have better transplant outcomes than “resistant” patients. Reasons for ESA unresponsiveness should be addressed in the clinical practice, with special attention to diagnosis and management of iron deficiencies.
MATERIALS AND METHODS
We performed a multicenter, observational, retrospective cohort study in 37 Spanish RT units with data collection between April and September 2011. The eligibility criteria were patients who received a kidney transplant in 2007, aged 18 years or older at the time of transplantation, with data available on Hb levels, and treated for anemia from the first month after transplantation. Patients with double kidney transplants, multiorgan transplants, and patients with thalassemia minor were excluded. To avoid selection bias, all consecutive patients fulfilling selection criteria in each center were included. The study protocol was approved by all ethics committees of centers, where it was required according to legislation in Spain.
We collected relevant medical history, transplant characteristics, and longitudinal laboratory data, which was assigned to study time points at 7 days (±2 days), 14 days (±2 days), 21 days (±2 days), 1 month (±1 week), 2 months (±1 week), 3 months (±1 week), 6 months (±2 weeks), 12 months (±4 weeks), 24 months (±4 weeks), and 36 months (±4 weeks) after transplantation. We also collected data on clinical course up to the time of data collection. Anemia was defined as Hb less than 11 g/dL and/or any ESA use or at least one RBC transfusion within the last 4 weeks. To explore the relationship between anemia management and transplant outcomes, patients were divided in subgroups defined by Hb levels and ESA use (Hb <11 g/dL without ESA [untreated anemics], Hb <11 g/dL with ESA [anemia nonresponders], Hb ≥11 g/dL with ESA [anemia responders], and Hb ≥11 g/dL without ESA [nonanemics]) and pretransplantation and early posttransplantation (first month) ESA use (no ESA pre-RT or post-RT, ESA pre-RT and post-RT, no ESA pre-RT and ESA post-RT, and ESA pre-RT and no ESA post-RT). A propensity score was also calculated using a multivariate logistic regression model to predict the probability of each individual of receiving ESA post-RT.
GFR was estimated using Modification of Diet in Renal Disease-4 equation (22). Differences between patients with and without anemia at 1 month were tested using Student’s t tests, Mann–Whitney tests, or chi-square tests, as applicable. We used univariate and multivariate Cox regression models to determine predictor factors for anemia at 1 month and to explore the association between variables related to anemia presence and/or management and the risk of death-censored graft loss or death from any cause (stepwise selection). The association between anemia and renal function over time was evaluated by a multivariate general lineal model repeated-measures analysis of variance. Statistical analyses were performed with the SAS package version 8.2. (SAS Institute, Cary, NC).
The study was conducted under the auspice of the Spanish Society of Transplantation. Writing assistance was funded by Amgen and provided by Dr. Neus Valveny from TFS Develop. Statistical analysis was performed by Marta Figueras, Xavier Núñez, and Andrew Shala from TFS Develop.
The authors wish to thank to all the ANEMIART Study participating investigators and centers (in alphabetical order): Alberto Rodríguez-Benot (Hospital Universitario Reina Sofía de Córdoba); Ana Ramos (Fundación Jiménez Díaz); Ana Sánchez-Fructuoso (Hospital Clínico San Carlos); Ángel Alonso (Complexo Hospitalario Universitario A Coruña); Antonio Franco (Hospital General Universitario de Alicante); Beatriz Sánchez (Hospital Universitario Puerta del Hierro); Carlos Gómez, Manuel Arias (Hospital Universitario Marqués de Valdecilla); Carlos Jiménez, Begoña Rivas (Hospital Universitario la Paz); Carmen de Gracia (Hospital Universitario Virgen de las Nieves); Daniel Serón (Hospital Universitari Vall d’Hebrón); Domingo Hernández (Hospital Regional Universitario Carlos Haya); Ernesto Gómez (Hospital Universitario Central de Asturias); Francisco Javier Paul (Hospital Universitario Miguel Servet); Francisco Llamas (Complejo Hospitalario Universitario de Albacete); Jaime Sánchez Plumed (Hospital Universitari i Politècnic La Fe); Jesús Bustamante (Hospital Universitario de Valladolid); José M. González (Hospital Universitario de Canarias); José M. Morales (Hospital Universitario 12 de Octubre); José M. Tabernero (Complejo Hospitalario de Salamanca); Josep M. Campistol (Hospital Clínic de Barcelona); Josep M. Grinyó (Hospital Universitari de Bellvitge); Juan J. Amenabar (Hospital de Cruces, Bilbao); Juan J. Cubero (Hospital Universitario Infanta Cristina); Julio E. Marco, Aina R. Obrador-Mulet (Hospital Universitari Son Espases); Julio Pascual, Marisa Mir (Hospital del Mar, Barcelona); Luis Guirado (Fundació Puigvert); Luis M. Pallardó (Hospital Universitario Dr. Peset); María A. Mazuecos (Hospital Universitario Puerta del Mar); María L. Rodríguez (Hospital General Universitario Gregorio Marañón); Mercedes Gil, Luisa Jimeno (Hospital Universitario Virgen de la Arrixaca); Miguel Á. Gentil (Hospital Universitario Virgen del Rocío); Miguel Á. Muñoz (Complejo Hospitalario de Toledo); and Pedro Errasti (Clínica Universitaria de Navarra).