Anemia Development and Cardiovascular Risk Management in Nonanemic Stage 3 Chronic Kidney Disease

Abstract

Background/Aim. There is little information on the development of anemia in the early stages of chronic kidney disease. The aim of this study was to analyze the onset of renal anemia in a cohort of initially nonanemic chronic kidney disease patients followed up in nephrology clinics. Methods. This epidemiological, prospective, three-year, multicenter study enrolled patients aged 18–78 years with stage 3 chronic kidney disease without anemia. Interim analysis was performed on the data collected during the first 12 months. Results. The study included 432 patients, average age 63.6 years (range 22–78 years, 70% male). The main etiologies of chronic kidney disease were glomerular (11.6%), interstitial (10.4%), vascular (29.4%), and diabetic (16.9%). The percentages of patients with comorbidities were 33.8% diabetes (2.5% type 1), 69% dyslipidemia, and 93% hypertension. During the first year, 12.4% of patients developed anemia. The chronic kidney disease progression rate was low: proteinuria was 0.46 ± 0.8 g/24 h at one year versus 0.67 ± 1.0 g/24 h at baseline. Diabetic patients showed a greater prevalence of previous cardiovascular events (50.0% vs. 24.5%) and worse control of some modifiable cardiovascular risk factors: smoking (13.4% vs. 8.6%), obesity (BMI > 30 kg/m², 33.6% vs. 25.3%), target blood pressure (<130/80 mmHg, 21.0% vs. 27.9%), and proteinuria (0.8 ± 1.1 vs. 0.6 ± 0.9 g/day). Conclusions. After one year, 12.4% of patients developed anemia. Diabetic patients had a higher cardiovascular risk and limited blood pressure control. The overall control of cardiovascular risk was unsatisfactory.

Keywords:

• chronic kidney disease
• anemia
• cardiovascular risk
• diabetes
• epidemiology

INTRODUCTION

Chronic kidney disease (CKD) is associated with high cardiovascular (CV) risk.^[1] Both classical and uremia-related risk factors contribute to increased morbidity. For this reason, current clinical guidelines place special emphasis on the importance of implementing a global CV risk reduction program, which may also reduce CKD progression.^[2] Although there is information about CV risk control in patients undergoing dialysis,^[3] scant data are available about the fulfillment of current guidelines at earlier CKD stages.

Anemia is one of the most common CV risk factors in renal patients and is a predictor of poor outcome (lower survival).^[4,5] Its prevalence increases as the glomerular filtration rate (GFR) decreases, especially below 60 mL/min.^[6] The most common onset is at a GFR of about 30 mL/min, although it may start earlier, at 45 mL/min, in diabetic patients.^[7] Impairment in the renal production of erythropoietin^[4,5,8,9] is the main reason for anemia in CKD patients, but other factors may also play a significant role. Retrospective studies at more advanced CKD stages have identified an association between the development of severe anemia and ethnic origin, diabetes, CKD severity, low serum iron levels, inflammation, and malnutrition.^[10] Among these factors, diabetes mellitus (DM) is one of the most frequently reported and important.^[7] However, most of the studies of the relationship between anemia and the CKD stage have a cross-sectional design, which does not allow the onset of anemia to be studied. Prospective studies involving stage 3 CKD patients are needed to identify those patients who would benefit most from early intervention against anemia.

The main objective of the study was to determine the percentage of stage 3 CKD patients who develop anemia, describe their characteristics, and identify the onset of anemia. The secondary objectives were to evaluate possible differences between diabetic and nondiabetic patients, describe the clinical evolution of the population and the factors associated with this evolution, describe the therapeutic management of several clinical conditions (i.e., CV risk, metabolic control, mineral metabolism, anemia, and hypertension [HT]) and guideline fulfillment in clinical practice, and determine whether anemia and poor prognosis are independent of GFR status. The present work presents the study design and an interim analysis of the primary endpoint and the first and third secondary objectives. All objectives will be included in the final analysis.

SUBJECTS AND METHODS

The NADIR-3 (“Study of Non-Anaemic Stage 3 CKD Patients Who Develop Renal Anaemia”) is an epidemiological, prospective, multicenter study conducted in Spain. It is promoted by the GEENDIAB (Spanish Group for the Study of Diabetic Nephropathy) within the Strategic Action Plan of the Spanish Society of Nephrology and has been approved by La Paz ethics committee.

Patients were included in the study if they were 18–78 years of age, had a Modification of Diet in Renal Disease [MDRD]-estimated GFR^[11] of 30–60 mL/min, and were without anemia. “Without anemia” was defined as two previous consecutive hemoglobin (Hb) concentrations of ≥11.5 g/dL in women or ≥12 g/dL in men aged > 70 years, or Hb ≥ 13.5 g/dL in men aged ≤ 70 years.^[7] Anemia onset (primary endpoint) was defined as the detection of an Hb level below the limits mentioned above that was confirmed by a second determination at least 15 days later. Patients with previous kidney transplant, treatment with erythropoiesis-stimulating agents (ESAs), who had received transfusions within three months of entry, or who had been treated with any immunosuppressive drug were excluded.

A systematic, consecutive sampling of patients was conducted between October 2005 and March 2006. The study is ongoing and follows each patient for three years or until the initiation of renal replacement therapy (RRT) through dialysis or renal transplant.

Clinical assessments were performed at the baseline (initial) visit and at follow-up visits conducted at six-month intervals, and included collection of clinical and biochemical data and therapeutic management information. The clinical and biochemical data included weight; height; blood pressure; hemogram; ferrokinetics; concentrations of folic acid, vitamin B12, creatinine (Cr), urea, albumin (Alb), lipids, glucose, glycosylated hemoglobin, calcium (Ca), phosphorus (P), parathyroid hormone, alkaline phosphatase (AP), C-reactive protein, and homocysteine; and proteinuria and proteinemia. Clinical events (e.g., hospitalization, death, RRT initiation) and anaemia development were recorded continuously during the study. For patients who developed anemia, the reason for the anemia diagnosis was investigated to identify non-CKD-related causes. Renal function was estimated according to the MDRD formula, and 24 h urine collection was also used.

Previous comorbidities and events during the study were based on clinical diagnosis of HT, dyslipidemia, and all other conditions included in the Charlson comorbidity index, which was validated previously for stage 5 CKD patients.^[12,13] DM and dyslipidemia were diagnosed according to the American Diabetes Association^[14] and National Cholesterol Education Program criteria.^[15] At each study visit, the presence of six modifiable CV risk factors was evaluated together with their degree of control according to current guidelines: HT (> 130/80 mmHg), proteinuria (>1.0 g/24 h), hyperlipidemia (concentration of total cholesterol [T-CHOL] >200 mg/dL or low-density lipoprotein cholesterol [LDL-C] >100 mg/dL), active smoking, malnutrition (Alb < 3.5 g/dL and body mass index [BMI] <20 kg/m²), and Ca × P product (>55 mg²/dL²). The presence of noncontrolled anemia (Hb < 11 g/dL) was also evaluated and considered as the seventh CV risk factor.^[16]

The estimated sample size needed for the primary objective was 427 patients to guarantee a precision of ± 4.6% (95% confidence interval) for an anemia incidence of 30% based on data from previous studies.^[6] This also accounted for a 10% dropout rate.

Categorical variables are presented as frequency (n) and percentage. Continuous variables are expressed as mean and standard deviation (SD) except for age (median and range). Student's t test or the Mann-Whitney U test was used to compare quantitative variables between subgroups. Pearson's chi-square test was used to compare qualitative variables. Data were analyzed using SPSS 12.0 statistical package (SPSS Inc., Chicago, Illinois, USA).

Experimental Investigation on Human Subjects

Institutional Review Board (IRB)/Ethics Committee approval was obtained.

RESULTS

Cohort Description

A total of 432 patients were included in the study, and these patients came from 27 nephrology services distributed across the country. Overall, 69.9% of patients were men and 30.1% women. The mean age was 63.6 years (range 22–78); 56% of the patients were more than 65 years of age. The mean creatinine clearance calculated according to the MDRD was 39.0 mL/min (42.9 mL/min using the Cockroft–Gault fomula). The main reasons for CKD (percentage of patients) were as follows: vascular (29.4%), DM (16.9%), interstitial (10.4%), adult polycystic kidney disease (6.9%), glomerulonephritis (11.6%), and unidentified (18.3%). About one-third (33.8%) of patients had DM (2.5% type 1). Table 1 displays the comorbidities in the overall group and in the subgroups of diabetic and nondiabetic patients. Ninety-three percent had HT, 98.5% of whom were receiving treatment (79.9% angiotensin converting enzyme inhibitors (ACEIs) or angiotensin receptor antagonists (ARAII)), and 10.9% both). The administration of ACEIs and ARAII was more frequent in diabetic patients (see Table 1). Diabetic patients had higher blood pressure (see Table 1). Sixty-nine percent of patients were hyperlipidemic, and 57.4% of these patients were receiving a hypolipidemic drug. Lipid targets of concentrations of LDL-C <100 mg/dL and triglycerides (TG) <500 mg/dL) were achieved in 40.1% of patients.

Table 1  Baseline comorbidities in the overall cohort and comparisons between diabetic and nondiabetic patients

	Total	DM	non-DM	p
Age, years	63.6 ± 11.7	66.7 ± 9.0	62.1 ± 12.6	<0.005
eGFR MDRD, mL/min	42.9 ± 8.4	43.7 ± 8.6	42.4 ± 8.2	0.142
Proteinuria, g/24 h	0.67 ± 1.0	0.8 ± 1.1	0.6 ± 0.9	0.081
Charlson index	3.21 ± 1.6	4.4 ± 1.8	2.6 ± 1.0	<0.001
Systolic BP, mmHg	134.5 ± 16.3	137.9 ± 17.5	132.9 ± 15.4	<0.005
Diastolic BP, mmHg	76.3 ± 9.6	74.8 ± 10.0	77.0 ± 9.4	0.054
Pulse pressure, mmHg	58.3 ± 14.8	63.1 ± 16.5	55.9 ± 13.4	<0.001
% BP-controlled (SBP <130 mmHg, DBP < 80 mmHg)	25.7	21.1	27.9	0.161
% Isolated systolic HT (SBP ≥ 130 mmHg, DBP <80 mmHg)	33.8	43.7	29	<0.005
% ACEI or ARAII	79.9	85.2	77.2	0.056
% Previous CV event	32.9	50	24.5	<0.001
% Smoker (active/ex-smoker)	10.2/39.1	13.4/35.9	8.6/40.7	0.16/0.39
% Alcohol (usual/periodic)	8.1/51.2	12.7/53.5	5.9/50.0	<0.05/0.56
% Obesity (BMI > 30 kg/m^2)	28.2	33.8	25.5	0.092
% LDL-C <100 mg/dL	40.1	53.2	33.8	<0.001
% HDL-C >60 mg/dL	23.5	18.2	26.4	0.09
% Triglycerides < 200 mg/dL	84.4	81.9	85.6	0.404
% Statins prescription	57.4	67.6	52.4	<0.005

Values are presented as mean ± SD and percentages.

Comparisons between DM and non-DM were performed using ANOVA or chi-square test, depending on the variable studied.

Abbreviations: DM = diabetes mellitus; eGFR MDRD= estimated glomerular filtration rate by Modified-Diet-Renal-Disease formula; BP = blood pressure; BMI = body mass index; LDL-C = low-density lipoprotein cholesterol; HDL-C = high-density lipoprotein cholesterol.

The mean analytical values for bone mineral disease (BMD-CKD) were corrected Ca, 9.4 ± 0.5 mg/L; P, 3.6 ± 0.6 mg/dL; Ca × P, 33.6 ± 6.5 mg²/dL²; and intact parathyroid hormone (i-PTH), 95.4 ± 65.2 pg/mL. The percentage of patients achieving the K/DOQI targets for Ca, P, and i-PTH was 54% in the overall group, 78.6% in diabetic patients, and 11.6% in nondiabetic patients. Phosphorus-chelating agents were administered to 5.8% of patients, and vitamin D or analogues were administered to 5.1%.

One patient initiated hemodialysis (HD), and three patients died during the first year of the study (mortality rate of 0.65 per 100 patient-years). At one year, the mean (± SD) measured GFR was 44.8 ± 15.2 mL/min and mean proteinuria was 0.46 ± 0.8 g/24 h.

Diabetic patients had a higher prevalence of previous CV events than nondiabetic patients (50% vs. 24.5%) and a trend toward poorer control of modifiable CV risk factors such as HT, proteinuria, smoking, and obesity at baseline (see Table 1). However, they displayed better control of dyslipidemia: 53.2% of patients in this subgroup achieved a LDL-C concentration of <100 mg/dL vs. 33.8% in nondiabetic patients (p < 0.001).

Anemia and Its Treatment

The mean baseline Hb concentrations were 13.3 g/dL in women, 14.4 g/dL in men >70 years, and 14.9 g/dL in men <70 years. Only 58.5% of patients presented with ferritin values <100 ng/mL or transferrin saturation index (TSI) < 20%, and 90.1% did not receive any iron supplementation. The mean ferritin concentration was 132 ng/mL, and the TSI was 30%. A lower percentage of patients (6.5%) received oral iron treatment (several formulations) at a mean dose of 1.77 ± 2.7 g/week. One patient was treated with intravenous iron.

During the first year of follow-up, 12.4% of patients developed anemia, which was primarily of renal etiology. Nine patients initiated anemia treatment; one patient received intravenous iron, and eight received ESAs.

Evolution of CV Risk Control

Table 2 shows the control of modifiable CV risk factors at baseline and after one year; these data are expressed as the percentage of patients who reached the treatment targets.

Table 2  Fulfillment of targets or treatment recommendations (if applicable) at baseline and one year later

	Baseline	One year	p
DM-specific targets
% DM patients with HbA1c <7.5%	70	75.5	0.42
% DM patients treated with antiplatelet agents	49.9	50.1	0.12
% DM patients with ACEI or ARAII	85.2	86.6	0.86
CKD targets
% Proteinuria <1 g/24 h	80.3	86.1	<0.05
BP targets
% BP controlled (SBP < 130 mmHg and DBP < 80 mmHg)	25.7	27.7	0.56
Lipid/metabolic targets
Not obese (BMI <30 kg/m^2)	71.8	69.8	0.59
% LDL-C 100 mg/dL treated with hypolipidemic agent	75.6	82.4	<0.05
% LDL-C <100 mg/dL	41.3	39.7	0.74
% Active smoker	10.4	8.9
% Never smoked	50.4	51.2	0.567
BMD targets
% Calcium × phosphorus <55 mg^2/dL^2	99.5	99.7	0.878
% Phosphorus >5.5 mg/dL without phosphate binders	0.2	0.3	0.637
Anemia targets
% Hb >11 g/dL	100	98.3	<0.01
% Hb <11d/dL without ESAs	0	0	1
% Ferritin <100 ng/mL or TSI <20% with iron treatment	47.3	51	0.378

Values are presented as percentages.

Targets are applicable only to diabetic patients and were analyzed only in the corresponding subgroup.

Abbreviations: BP = blood pressure; LDL-C = low-density lipoprotein cholesterol; HbA1c = glycosylated hemoglobin; BMD = bone mineral disease; Hb = hemoglobin; ESAs = erythropoiesis-stimulating agents; TSI = transferrin saturation index; ACEI = angiotensin converting enzyme inhibitor; ARAII = angiotensin receptor antagonists.

Analysis of the combined presence of six basic risk factors described in the literature showed that 85.7% of patients had at least one uncontrolled risk factor at the baseline visit. Table 3 shows some characteristics in subgroups of patients defined by their number of uncontrolled risk factors. The variables associated with an increased number of uncontrolled risk factors were younger age, lack of previous CV event, and higher serum Cr concentration.

Table 3  Characteristics of patients in subgroups defined by the number of uncontrolled cardiovascular risk factors at baseline

Total uncontrolled RF	0–1 (159)	2 (124)	≥ 3(45)	p
Age, years	65.49 ± 10.01	63.44 ± 11.14	56.76 ± 13.38	<0.001
Serum creatinine, mg/dl	1.78 ± 0.43	1.83 ± 0.48	1.99 ± 0.46	<0.05
GFR, mL/min	42.83 ± 8.25	42.55 ± 8.2	43.07 ± 8.67	NS
% Men	69.2%	66.9%	71.1%	NS
% Diabetes	33.3%	33.9%	31.1%	NS
% Previous CV event	17.7%	37.7%	37.8%	<0.005

*Significant differences between the selected subgroup and the other groups.

†Significant differences between the two selected subgroups.

Uncontrolled cardiovascular risk factors are based on score including one point for each of the following items: SBP/DBP > 130/80 mmHg; proteinuria > 1 g/24 h, T-CHOL > 200 mg/dL, or LDL-C > 100 mg/dL; malnutrition; active smoking or Ca × P > 55 mg²/dL² at baseline.

Values are presented as mean ± SD or number (percentage).

Abbreviations: GFR = glomerular filtration rate, LDL-C = low-density lipoprotein cholesterol, RF = risk factor, CV = cardiovascular, NS = non significan.t

DISCUSSION

The prevalence of CKD at stage 3 or higher^[17,18] is 6.8% of the general population in Spain.^[18] Anemia is a common complication of CKD, and the risk of developing anemia is higher in individuals with low GFR.^[6,19,20] A low Hb concentration diminishes the patient's quality of life and decreases functional status, both of which add to the economic cost associated with the treatment of CKD. Anaemia also seems to be a prognosis factor for poor outcomes in advanced CKD stages (5D).^[21,22] Some epidemiological studies have reported a 5% prevalence of anemia in patients with stage 3 CKD and 45% in stage 4.^[6] Other Spanish studies conducted in the nephrology setting have reported anemia in one of every two patients with stages 3 and 4 CKD.^[23] However, these data come from cross-sectional studies or from post hoc analysis of large databases that were not intended to investigate anemia management. Our currently ongoing study is the first prospective follow-up of a nonanemic stage 3 CKD cohort of patients and is intended to record the timing of anemia onset. During the first year, 12.4% of patients developed anemia despite the low rate of CKD progression, as defined by GFR and proteinuria values. The description of the characteristics of anemic patients and the analysis of factors associated with anemia onset will be performed at the end of the study.

CKD patients who are followed up at nephrology clinics usually display high CV comorbidity,^[24,25] poor prognosis with progression to RRT in 25% of cases at five years in some cohorts,^[24] and mortality rates as high as 45%. The comorbidities in our cohort were more frequent than in the general population but were less prevalent than in other stages 3 and 4 CKD populations described previously.^[16] This may explain the better evolution during the first year of follow-up in our cohort. Another possible explanation may be the “Hawthorne effect”—a positive effect on all patients included in clinical studies (including observational–epidemiological studies) because of the greater systematic attention, which improves the outcomes. The overall CKD progression rate in our cohort was lower than 1 mL/min per year, proteinuria was reduced, and mortality and RRT initiation rates were very low compared with other studies of stages 3 and 4 CKD patients. In a previous report from stages 3 and 4 CKD patients not selected according to their anemia status, the annual mortality rate was 3% and RRT initiation was 4% for the stage 3 subgroup.^[23] In that study, patients were three years older, 61.6% had stage 4 CKD, and the comorbidity was similar except for an anemia prevalence of 51.3%, according to Kidney Disease: Improving Global Outcomes (KDIGO) criteria, and 30.5%, according to European best practice guidelines (EBPG). The absence of anemia seems to be related to a lower risk profile, although this hypothesis must be verified in studies designed specifically to test it, as in our present study.

Stage 3 diabetic patients have a higher prevalence of previous CV events, and for this reason, their CV risk factors should be controlled more strictly. The blood pressure values obtained at baseline were better than those reported in patients without CKD in previous studies conducted in primary care settings.^[25] However, despite receiving ACEIs or ARAII in 9 of 10 patients, our diabetic patients still had poorer control than the nondiabetic patients.

Lifestyle habit-dependent risk factors, such as obesity, smoking, and alcohol intake, were also more prevalent in diabetic patients. The control of glycosylated Hb the diabetic subgroup was better than in the general population,^[26] and lipid control was even better than in our nondiabetic patients. Previous studies have suggested that some therapeutic “nihilism” may be important in these patients,^[23] although in our cohort, the percentage of patients not being treated with statins, antihypertensives, or antiplatelet agents was very low.

In the past 10 years, many international guidelines and recommendations have been published about the optimum therapeutic management of CV risk,^[27] anemia,^[7] and BMD.^[28] When we looked at the main modifiable CV risk factors, we found an acceptable degree of control at baseline and that several aspects had improved at the end of the first year of follow-up. However, the positive results reported in the separate analysis of risk factor control were worsened markedly in the conjoint analysis. Only 1% of patients fulfilled all the targets (i.e., HT, dyslipidemia, DM [if applicable], and BMD) simultaneously and throughout the whole follow-up period. This finding justifies the strong recommendation for a closer clinical follow-up and increased drug adjustments at each visit to try to reduce the duration of “out-of-control” periods for each patient.

Only the CKD severity and the presence of a previous CV event were associated with poorer control of the seven most relevant risk factors identified in previous reports.^[13] The number of targets achieved has prognostic value in CKD patients at more advanced stages and may also have prognostic value at earlier stages.^[29]

In conclusion, our study allows us to identify the timing of anemia onset and the factors associated with its development in CKD patients. Our current data provide us with valuable information about areas of improvement for reducing mortality risk in nonanemic stage 3 CKD patients, though the overall control of CV risk in our patients is far from optimal.

ACKNOWLEDGMENTS

This study has been conducted by the above-signed authors and the GEENDIAB-SEN Group (in alphabetical order): Dr. Vicente Álvarez Chivas, H. de La Princesa (Madrid); Dr. Pedro Aranda, H. Carlos Haya (Málaga); Dr. Jesús Mª Arteaga Coloma, H. de Navarra; Dra. Josefa Borrego, Complejo Hosp. de Jaén (Jaen); Dr. Jesús Bustamante Bustamante, H. Clínico de Valladolid (Valladolid); Dra. Francisca Calero, Fundació Puigvert (Barcelona); Dra. Camino, H. Central de Asturias (Oviedo); Dr. Aleix Cases Amenós, H. Clínic (Barcelona); Dr. José Conde, H. de Toledo (Toledo); Dr. Fernando de Álvaro, H. La Paz (Madrid); Dr. Francisco Fernández Vega, H. Central de Asturias (Oviedo); Dr. Florencio García Martín, H. 12 de Octubre (Madrid); Dr. Joan Gascó González, H. Son Llatzer (Palma de Mallorca); Dr. José Luis Gorriz Teruel, H. Dr. Peset (Valencia); Dr. Manuel Granda Rodríguez, H. Virgen Blanca (León); Dr. José María Graña Fandos, H. de La Rivera (Valencia); Dr. Enrique Gruss Vergara, H. U. Fundación Alcorcón (Madrid). Dr. Pablo Íñigo Gil, H. Clínico Universitario Lozano Blesa (Zaragoza); Dra. Dolores Lorenzo, H. Juan Canalejo (A Coruña); Dr. Rafael Marín Iranzo, H. Central de Asturias (Oviedo); Dr. Alberto Martínez Castelao, H. de Bellvitge (Barcelona); Dra. Isabel Martínez Fernández, H. de Galdakao (Vizcaya); and Dr. Javier Nieto Iglesias, H. Gral. de Ciudad Real (Ciudad Real).

This study was supported financially by Amgen S.A., which also provided editorial assistance for the preparation of this manuscript. The authors have previously attended advisory boards for Amgen S.A. This study has been designed and performed by the authors and the GEENDIAB Group of the Spanish Society of Nephrology.