- Email: [email protected]
Anaemia, independent of chronic kidney disease, predicts all-cause and cardiovascular mortality in type 2 diabetic patients
Atherosclerosis 210 (2010) 575–580
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Anaemia, independent of chronic kidney disease, predicts all-cause and cardiovascular mortality in type 2 diabetic patients Giacomo Zoppini a,1 , Giovanni Targher a,∗,1 , Michel Chonchol b , Carlo Negri a , Vincenzo Stoico a , Isabella Pichiri a , Giuseppe Lippi c , Michele Muggeo a , Enzo Bonora a a Section of Endocrinology, Department of Biomedical and Surgical Sciences, University of Verona, Ospedale Civile Maggiore, Piazzale Stefani, 1, 37126 Verona, Italy b Division of Renal Diseases and Hypertension, University of Colorado Denver Health Sciences Center, Aurora, CO, USA c Section of Clinical Chemistry, Department of Biomedical and Morphological Sciences, University of Verona, Verona, Italy
a r t i c l e
i n f o
Article history: Received 22 September 2009 Received in revised form 2 December 2009 Accepted 2 December 2009 Available online 11 December 2009 Keywords: Anaemia Chronic kidney disease Type 2 diabetes Mortality Epidemiology
a b s t r a c t Objective: There is limited and controversial information on whether anaemia is a risk factor for cardiovascular mortality in type 2 diabetes, and whether this risk is modified by the presence of chronic kidney disease (CKD). We assessed the predictive role of lower hemoglobin concentrations on all-cause and cardiovascular mortality in a cohort of type 2 diabetic individuals. Methods: The cohort included 1153 type 2 diabetic outpatients, who were followed for a mean period of 4.9 years. The independent association of anaemia (i.e., hemoglobin <120 g/l in women and <130 g/l in men) with all-cause and cardiovascular mortality was evaluated by Cox proportional hazards regression models and adjusted for several potential confounders, including kidney function measures. Results: During follow-up, 166 (14.4%) patients died, 42.2% (n = 70) of them from cardiovascular causes. In univariate analysis, anaemia was associated with increased risk of all-cause (hazard ratio HR 2.62, 95% confidence intervals 1.90–3.60, p < 0.001) and cardiovascular mortality (HR 2.70, 1.67–4.37, p < 0.001). After adjustment for age, sex, body mass index, smoking, hypertension, dyslipidemia, diabetes duration, hemoglobin A1c, medication use (hypoglycemic, anti-hypertensive, lipid-lowering and anti-platelet drugs) and kidney function measures, the association of anaemia with all-cause (adjusted HR 2.11, 1.32–3.35, p = 0.002) and cardiovascular mortality (adjusted HR 2.23, 1.12–4.39, p = 0.020) remained statistically significant. Conclusions: Anaemia is associated with increased risk of all-cause and cardiovascular mortality in type 2 diabetic individuals, independently of the presence of CKD and other potential confounders. The advantage to treat anaemia in type 2 diabetes for reducing the risk of adverse cardiovascular outcomes remains to be demonstrated. © 2009 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Several epidemiologic studies have consistently shown that anaemia and chronic kidney disease (CKD) are both important risk factors of all-cause and cardiovascular mortality in the general population and in some groups of high-risk patients [1–5]. However, the information on the relationship among anaemia, CKD and cardiovascular outcomes in people with type 2 diabetes is limited and controversial, even though anaemia is common in type 2 diabetic patients with CKD. In fact, anaemia is often more severe and occurs at an earlier stage of CKD in patients with
∗ Corresponding author. Tel.: +39 045 8123748; fax: +39 045 8027314. E-mail address: [email protected] (G. Targher). 1 These authors contributed equally to this work. 0021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2009.12.008
type 2 diabetes compared with those with other causes of kidney disease [5]. In a recent study Vlagopoulos et al. [6] showed that anaemia, as defined by low levels of hematocrit (i.e., <36% in women and <39% in men), was a significant risk factor for cardiovascular disease and all-cause mortality primarily in diabetic patients who also had CKD (defined as an estimated glomerular filtration rate <60 ml/min/1.73 m2 ), whereas anaemia was not a risk factor for any outcome in those with normal or near-normal kidney function [6]. In contrast, Tong et al. reported that in a cohort of Chinese type 2 diabetic patients, who were followed for ∼3 years, the risk of future cardiovascular events increased progressively in presence of decreasing levels of hematocrit only in those without declining kidney function [7]. The aim of this study was to examine whether anaemia is associated with increased risk of all-cause and cardiovascular mortality
576
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
in a cohort of type 2 diabetes patients, and whether this risk is modified by the presence of CKD.
2. Methods The study was performed within the frame of the Verona Diabetes Study, an observational longitudinal study on chronic complications in type 2 diabetic outpatients attending the diabetes clinic at the University Hospital of Verona [8,9]. Data included in this analysis are based upon the cohort of Caucasian type 2 diabetic outpatients (n = 1153), who were recruited over the period of January 2000 to January 2002 and then followed-up until September 30, 2007, after excluding (i) patients who had a history of malignancy, chronic obstructive pulmonary disease, end-stage renal disease or cardiovascular disease (defined as angina, myocardial infarction, revascularization procedures and stroke), and (ii) those who had incomplete clinical and biochemical data for analysis. Type 2 diabetes was established when diagnosis was made after the age of 35 years, irrespective of treatment, or when the disease was treated with diet or oral hypoglycemic agents, irrespective of age at diagnosis. More details about the study design and recruitment methods have been reported elsewhere [10]. Briefly, the 1153 participants included in this analysis represented approximately 30% of the whole sample of type 2 diabetic outpatients (n = 3727), who regularly attended our diabetes clinic during years 2000–2002 and who had complete laboratory data for analysis after excluding patients with the above-mentioned comorbid conditions (n = 989, 26.5%). Notably, baseline demographics (∼50% male; mean age: 67 years vs. 66 years; mean body mass index: 27.9 kg/m2 vs. 28.2 kg/m2 ; mean diabetes duration: 16.0 years vs. 15.4 years and mean hemoglobin A1c 7.5% vs. 7.6%, respectively) as well as the overall crude rates of all-cause (14.4% vs. 14%) and cardiovascular (6.1% vs. 6.5%) mortality were not significantly different between the 1153 participants of the study and those with missing data for hemoglobin or other laboratory variables (n = 1585). All participants were periodically seen (every 3–6 months) for routine medical examinations of glycemic control and chronic complications of diabetes. The local ethics committee approved the study protocol. All participants gave their informed consent. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters. Blood pressure was measured with a standard mercury manometer. Information on comorbid conditions, current use of medications and smoking history was obtained in all patients by interviews during medical examinations. In all participants venous blood was withdrawn in the morning after an overnight fast for standard biochemical work-up. Hemoglobin concentration was determined by a photometrical technique on the fully automated hematological analyzer ADVIA 120-TM (Bayer Diagnostics, Milan, Italy). Serum creatinine (measured by using the Jaffé method – kinetic alkaline picrate assay), lipids and other biochemical blood measurements were determined by automatic colorimetric methods (DAX 96, Bayer Diagnostics, Milan, Italy). LDL-cholesterol was calculated by Friedewald’s equation, except in those with triglycerides exceeding 4.55 mmol/l (n = 36). Hemoglobin A1c was measured by a high-performance liquid chromatography analyzer (BioRad Diamat, Milan, Italy) and the upper limit of normality was 5.6%. Anaemia was defined as hemoglobin concentrations <120 g/l in women and <130 g/l in men, according to the World Health Organization criteria. Glomerular filtration rate was estimated (eGFR) from the four-variable Modification of Diet in Renal Disease (MDRD) equation [11] as follows: eGFR = 186 × (serum creatinine−1.154 ) × (age−0.203 ) × 1.212 (if black) × 0.742 (if female).
Urinary albumin excretion was measured from an early morning urine sample on at least three consecutive occasions, within a period of 4–6 months, as the albumin/creatinine ratio (ACR) by an immuno-nephelometric method. Microalbuminuria was defined as an ACR of 30–299 mg/g and macroalbuminuria as an ACR of ≥300 mg/g [12]. The presence of CKD was diagnosed when a patient had either eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria (i.e., micro- or macroalbuminuria) irrespective of eGFR values [11,12]. Hypertension was defined as a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg or current use of any treatment with anti-hypertensive medications. Information regarding specific classes of anti-hypertensive drugs (e.g., ACE-inhibitors and angiotensin receptor blockers) was not currently available. Atherogenic dyslipidemia was diagnosed when plasma triglycerides were ≥1.70 mmol/l and/or HDL-cholesterol was <1.04 mmol/l or when patients were taking lipid-lowering agents [12]. Vital status on September 30, 2007 was ascertained for all participants (n = 1153) by examining the database of the Social Health Unit of the Veneto Region, which includes all records of mortality occurred within the Veneto Region as well as the specific causes of death. Causes of death were identified in 100% of subjects. Trained nosologists coded death certificates using the International Classification of Diseases, Ninth Revision (ICD-9). Deaths were attributed to cardiovascular diseases when ICD-9 codes were 390–459. A selected sample of death certificates was reviewed manually to validate the process. 2.1. Statistical analysis Data are presented as means ± SD or proportions. Skewed variables were logarithmically transformed to improve normality prior to analyses. The unpaired t-test and the chi-squared test with Yates’s correction for continuity (for categorical variables) were used to compare the baseline clinical characteristics of participants stratified by the presence or absence of anaemia, defined as hemoglobin <120 g/l in women and <130 g/l in men. Univariate survival analysis stratified by anaemia and CKD status was performed by the Kaplan–Meier analysis, and the overall significance was calculated by the log-rank test. Cox regression analysis was used to study the effect of anaemia on the risk of all-cause and cardiovascular mortality after adjustment for potential confounders. Four forced entry multivariate regression models were performed. Hemoglobin was included as either continuous or dichotomous variable in these models. The first multivariate regression model was adjusted for age (years), gender (male vs. female) and BMI (kg/m2 ); the second model for age (years), gender (male vs. female), BMI (kg/m2 ), diabetes duration (years), hemoglobin A1c (%), hypertension (yes/no; see definition above), atherogenic dyslipidemia (yes/no; see definition above), LDL-cholesterol (mmol/l) and current use of medications (anti-platelet and hypoglycemic drugs) (yes/no); finally, the third and fourth regression models were further adjusted for eGFR and albuminuria, which were included as either categorical (i.e., eGFR <60 ml/min/1.73 m2 vs. ≥60 ml/min/1.73 m2 and ACR <30 mg/g vs. ≥30 mg/g, respectively) or continuous measures. The covariates included in multivariate regression models were chosen as potential confounders based on their biological plausibility or statistical association with mortality in univariate analysis. Results are presented as hazard ratios (HRs) with 95% confidence intervals (CI) and statistical significance was evaluated by the likelihood-ratio test. HRs for continuous variables were computed for each SD change. Statistical analyses were performed with statistical package SPSS 14.0. p-Values <0.05 were considered statistically significant.
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
577
Table 1 Baseline characteristics of type 2 diabetic patients stratified by anaemia status.
Sex (M/F) Age (years) Diabetes duration (years) Body mass index (kg/m2 ) Current smokers (%) Hypertension (%) Hemoglobin A1c (%) LDL-cholesterol (mmol/l) Triglycerides (mmol/l) HDL-cholesterol (mmol/l) Hemoglobin (g/l) eGFR (ml/min/1.73 m2 ) Micro- or macroalbuminuria (%) Chronic kidney disease (%)a Oral hypoglycaemic drugs (%) Insulin therapy (%) Anti-hypertensive drugs (%) Lipid-lowering drugs (%) Anti-platelet drugs (%)
No anaemia (n = 941)
Anaemia (n = 212)
p
477/464 67 ± 12 16 ± 9 28.1 ± 5.0 12.7 77 7.4 ± 1.5 3.41 ± 0.9 1.63 ± 1.1 1.41 ± 0.4 144.7 ± 12 71.6 ± 18 23.7 45.9 73.2 32.1 70.0 40.3 38.9
101/111 67 ± 15 15 ± 9 27.2 ± 4.5 29.8 82 7.6 ± 1.6 3.10 ± 0.9 1.71 ± 0.9 1.40 ± 0.4 106.1 ± 21 61.1 ± 24 32.2 67.9 65.6 50.5 73.1 33.5 42.5
0.45 0.52 0.01 0.03 0.001 0.12 0.06 <0.001 0.85 0.24 ND <0.001 0.03 <0.001 <0.001 <0.001 0.41 0.07 0.36
Cohort size, n = 1153. Anaemia was defined as hemoglobin <120 g/l in women and <130 g/l in men. Data are expressed as means ± SD or proportions. p-Values refer to t-test or the 2 test (for categorical variables). LDL, low density lipoprotein; HDL, high density lipoprotein; eGFR, estimated glomerular filtration rate; ND, not determined. a Chronic kidney disease was defined as eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria.
3. Results Overall, anaemia was present in 18.4% (n = 212; 52.4% female) of the whole cohort; mean [SD] hemoglobin: 106.1±21 g/l, median [interquartile range, IQR]: 113.5 [99–120] g/l in all anaemic participants; mean [SD]:101.2 ± 21 g/l, median [IQR]: 110.5 [87–116] g/l in anaemic women, and mean [SD]: 111.5 ± 21 g/l, median [IQR]: 119.0 [106–128] g/l in anaemic men. The baseline characteristics of participants stratified by anaemia status are shown in Table 1. Compared with those with normal hemoglobin levels, participants with anaemia had a greater frequency of CKD (defined as eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria), lower BMI, lower LDL-cholesterol, and tended to have higher hemoglobin A1c and greater prevalence of hypertension. In addition, they also had a shorter duration of diabetes and were more likely to be smokers and to be treated with insulin therapy compared with those without anaemia. Age, gender, triglycerides, HDL-cholesterol and proportions using lipidlowering, anti-platelet or anti-hypertensive medications were not significantly different between the groups. During the follow-up (mean ± SD: 4.9 ± 1.0, range: 1–6 years), 166 (14.4%) participants died, 70 (42.2%) of them from cardiovascular causes. The crude rates for all-cause and cardiovascular mortality were 58 (27.4%) and 26 (12.3%) in anaemic patients, and 108 (11.5%) and 44 (4.7%) in those without anaemia. Fig. 1 shows the results of the Kaplan–Meier survival analysis for all-cause mortality in participants stratified by anaemia and CKD status. Compared with those with no anaemia and no CKD (included as the referent group), participants with both anaemia and CKD had the highest risk of mortality, whereas those with anaemia or CKD alone had an intermediate risk (p < 0.001 by the log-rank test). Almost identical results were obtained for cardiovascular mortality (data not shown). The independent association of lower hemoglobin concentrations with all-cause and cardiovascular mortality was tested by Cox regression analysis (Table 2). In unadjusted regression models, lower hemoglobin was associated with a higher risk of all-cause and cardiovascular mortality when it was included as either continuous (for each SD decrement, i.e., 21 g/l) or dichotomous variable. Of note, the association of lower hemoglobin with all-cause and cardiovascular mortality remained significant even after adjustment for a broad spectrum of potential confounders, including the
presence of kidney dysfunction (Table 2, model 1–4). Others independent predictors of all-cause and cardiovascular mortality were older age, male sex, smoking history and presence of CKD (data not shown). We tested for a formal interaction between anaemia and eGFR <60 ml/min/1.73 m2 and between anaemia and abnormal albuminuria for each of the outcomes by including their crossproduct in the fully adjusted multivariate regression model. The anaemia × eGFR <60 ml/min/1.73 m2 interaction term (p = 0.084 and p = 0.395 for all-cause and cardiovascular mortality, respectively) and the anaemia × abnormal albuminuria interaction term (p = 0.385 and p = 0.174 for all-cause and cardiovascular mortality, respectively) were not statistically significant for any of the outcomes, except for a borderline significance for the anaemia–eGFR interaction term for all-cause mortality (p = 0.084). As shown in Table 3, we also examined the association of sexspecific quintiles of hemoglobin concentration with the risk of all-cause and cardiovascular mortality. The unadjusted HRs for allcause and cardiovascular mortality progressively increased with decreasing quintiles of hemoglobin concentration. After adjust-
Fig. 1. Kaplan–Meier survival analysis for all-cause mortality in 1153 type 2 diabetic patients stratified by anaemia and CKD status (no anaemia and no CKD as the reference category). The overall statistical significance was estimated by the log-rank test (p < 0.001).
578
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
Table 2 Cox proportional hazards models of lower hemoglobin levels for all-cause and cardiovascular mortality in type 2 diabetic patients. 1-SD decrement in hemoglobin
p
Anaemia (yes/no)
p
All-cause mortality Unadjusted HR (95% CI)
1.37 (1.20–1.53)
<0.001
2.62 (1.90–3.60)
<0.001
Adjusted HR (95% CI) Model 1 Model 2 Model 3 Model 4
1.45 (1.26–1.67) 1.41 (1.22–1.61) 1.24 (1.05–1.47) 1.18 (0.99–1.45)
<0.001 <0.001 0.010 0.07
2.39 (1.69–3.39) 2.18 (1.52–3.13) 2.11 (1.32–3.35) 1.91 (1.17–3.07)
<0.001 <0.001 0.002 0.009
Cardiovascular mortality Unadjusted HR (95% CI)
1.39 (1.18–1.67)
<0.001
2.70 (1.67–4.37)
<0.001
Adjusted HR (95% CI) Model 1 Model 2 Model 3 Model 4
1.45 (1.18–1.79) 1.37 (1.11–1.72) 1.35 (1.04–1.80) 1.30 (1.01–1.72)
0.001 0.004 0.032 0.050
2.31 (1.36–3.91) 2.10 (1.19–3.57) 2.23 (1.12–4.39) 1.89 (1.03–3.76)
0.002 0.009 0.020 0.043
Cohort size, n = 1153. Results are expressed as hazard ratios (HRs) and 95% confidence intervals (in parenthesis). Model 1: adjusted for age, sex and body mass index. Model 2: model 1 plus further adjustment for diabetes duration, hemoglobin A1c, smoking history, hypertension, atherogenic dyslipidemia, LDL-cholesterol, and current use of medications. Model 3: model 2 plus further adjustment for eGFR (<60 ml/min/1.73 m2 vs. >60 ml/min/1.73 m2 ) and albuminuria (ACR <30 mg/g vs. >30 mg/g) included as categorical variables. Model 4: model 2 plus further adjustment for eGFR and albuminuria included as continuous variables. In both the unadjusted regression models and the four multivariate regression models hemoglobin was included as either continuous (for each SD decrement) or dichotomous variable (i.e., hemoglobin <120 g/l in women and <130 g/l in men).
Table 3 Cox proportional hazards models of sex-specific quintiles of hemoglobin concentrations for all-cause and cardiovascular mortality in type 2 diabetic patients. I quintile <121 g/l W <132 g/l M
II quintile 121–131 g/l W 132–145 g/l M
III quintile 132–139 g/l W 146–151 g/l M
IV quintile 140–145 g/l W 152–157 g/l M
V quintile >146 g/l W >157 g/l M
All-cause mortality Unadjusted HR (95% CI) Fully adjusted HR (95% CI)
3.9 (2.3–6.5) 2.2 (1.3–3.9)
1.9 (1.1–3.4) 1.2 (0.7–2.1)
1.4 (0.9–2.5) 1.1 (0.5–1.9)
1.2 (0.8–2.4) 0.8 (0.4–1.7)
Ref. Ref.
<0.001 0.018
Cardiovascular mortality Unadjusted HR (95% CI) Fully adjusted HR (95% CI)
3.9 (1.8–8.2) 2.1 (1.2–5.3)
2.0 (1.1–4.4) 1.4 (0.8–3.6)
1.3 (0.8–3.2) 1.0 (0.6–2.7)
1.0 (0.5–2.5) 0.7 (0.4–1.5)
Ref. Ref.
<0.001 0.025
p for linear trend
Cohort size, n = 1153. Results are expressed as hazard ratios (HRs) and 95% confidence intervals (in parenthesis). W = women; M = men; Ref. = reference category. Fully adjusted regression models: adjustment for age, sex, body mass index, diabetes duration, hemoglobin A1c, smoking history, hypertension, atherogenic dyslipidemia, LDL-cholesterol, medication use, eGFR (<60 ml/min/1.73 m2 vs. >60 ml/min/1.73 m2 ) and albuminuria (ACR <30 mg/g vs. >30 mg/g).
ment for age, sex, BMI, smoking, hypertension, dyslipidemia, diabetes duration, hemoglobin A1c, medication use and kidney function measures, the risk of all-cause and cardiovascular mortality across hemoglobin quintiles was consistently attenuated and it was more pronounced among those belonging to the lowest hemoglobin quintile. Fig. 2 shows the age- and sex-adjusted association of anaemia with the risk all-cause mortality in subgroup analyses. Notably, anaemia was significantly associated with an increased risk of
Fig. 2. Subgroup analyses of the risk of all-cause mortality by anaemia. Hazard ratios (95% confidence intervals) are adjusted for age and sex. p-Values for the interaction term of anaemia with each stratification variable, which were observed in an ageand sex-adjusted regression model, are also reported.
mortality in all subgroups of participants, independently of their baseline values of BMI, hemoglobin A1c, eGFR and albuminuria. Almost identical results were observed for cardiovascular mortality (data not shown). 4. Discussion The major finding of this study is that anaemia, defined as hemoglobin <120 g/l in women and <130 g/l in men, is strongly associated with an increased risk of all-cause and cardiovascular mortality in a large outpatient cohort of type 2 diabetic individuals. The presence of anaemia roughly doubled the risk of dying, mainly from cardiovascular causes. Moreover, the risk of cardiovascular mortality increased approximately 40% for each SD decrement in hemoglobin concentration. Notably, the association of lower hemoglobin concentrations with the risk of all-cause and cardiovascular mortality was independent of a broad number of potential confounders, including traditional cardiovascular risk factors, diabetes-related variables, kidney function measures (eGFR and albuminuria) and use of medications, i.e., all factors potentially correlated also to raised hemoglobin concentrations [6,13–15]. Several prospective studies have shown that anaemia and CKD are both important risk factors of all-cause and cardiovascular mortality in the general population and in some groups of high-risk patients [1–5]. At present, however, the available information on the association among anaemia, CKD and adverse cardiovascular outcomes in people with type 2 diabetes is quantitatively limited and controversial.
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
In a small retrospective study of 508 type 2 diabetic patients with established nephropathy, it was found that anaemia strongly predicted all-cause mortality [16]. In a recent pooled analysis from four community-based studies of 3015 US diabetic individuals, who were followed for a median period of 8.6 years, anaemia was a risk factor for fatal and non-fatal cardiovascular events primarily in those who also had CKD. The combination of anaemia and CKD conferred a synergistic risk for adverse cardiovascular outcomes compared with each factor risk alone. In contrast, anaemia was not a significant risk factor for any outcome in those with normal or near-normal kidney function [6]. It is important to emphasize that the authors of this study diagnosed anaemia extrapolating hemoglobin concentrations from those of hematocrit, which were measured in four different laboratories, and did not adjust their results for some potentially important confounders, such as diabetes duration, hemoglobin A1c and medication use [6]. Moreover, the diagnosis of CKD was based only on eGFR values, whereas information on albuminuria was not available for diagnosing CKD as recommended by the recent National Kidney Foundation guidelines [11]. Conversely, in a cohort of 3983 Chinese type 2 diabetic patients who were followed for a median period of 3 years, lower levels of hematocrit were associated with increased risk of all-cause mortality and non-fatal cardiovascular events only in those without CKD [7]. In that study, however, the study cohort was hospital based, making selection bias a potential confounding factor. Moreover, the specific causes of death were not analysed, and the results were not adjusted for potentially important confounders, such as the use of hypoglycemic, lipid-lowering and anti-platelet drugs [7]. In the present study, we found that the association between anaemia and increased risk of all-cause and cardiovascular mortality was only partially modified by the presence of CKD. Although a large part of the excess hazards of hemoglobin for all-cause and cardiovascular mortality was likely explained by the presence of kidney dysfunction (from ∼20% to 50% when comparing model 4 to model 2 of Table 2), it appeared that anaemia is associated with increased risk of all-cause and cardiovascular mortality even in individuals with normal eGFR and normo-albuminuria but that this association may be stronger in those who also had CKD, as also confirmed in our subgroup analyses (Fig. 2). Interestingly, the anaemia–CKD interaction terms were not statistically significant for any of the study outcomes, except for a borderline significance for the anaemia–eGFR interaction term for all-cause mortality (p = 0.084). When we examined the association of sex-specific quintiles of hemoglobin concentration with mortality, the unadjusted HRs for all-cause and cardiovascular mortality increased progressively with decreasing quintiles of hemoglobin concentration. After adjustment for multiple potential confounders, including also kidney function measures, the risk of all-cause and cardiovascular mortality across hemoglobin quintiles was largely attenuated and it was more pronounced among those belonging to the lowest hemoglobin quintile (Table 3). Although the p-values for the linear trend in this analysis were statistically significant and suggested a possible linearity of the hemoglobin concentration with mortality risks, a potential threshold below the IIIrd quintile of hemoglobin cannot be definitely excluded from these data. Anaemia might be just a marker of underlying established risk factors or important comorbidities. Notably, in the present study we excluded patients who had a history of malignancy, severe chronic obstructive pulmonary disease, end-stage renal disease or cardiovascular disease. Moreover, we found that lower hemoglobin concentration was associated with increased risk of all-cause and cardiovascular mortality independently of several established risk factors and potential confounders. Thus, it is conceivable that anaemia might confer an excess risk over and above the risk attributable to underlying established risk factors. The recognition of the biological mechanism(s) linking anaemia
579
to cardiovascular mortality is beyond the scope of this study. However, some putative biological mechanisms might be mentioned. It is known that anaemia may result in many potentially adverse effects on the cardiovascular system, such as increased cardiac output and increased heart rate. Over the long term, adaptations that initially increase cardiac output lead to left ventricular enlargement and eccentric left ventricular hypertrophy. Thus, it is likely that anaemic patients with left ventricular hypertrophy or subclinical ischemic heart disease might be at higher risk of developing symptomatic cardiac disease [4–6,17–22]. Anaemia may activate the sympathetic nervous system, stimulate the renin–angiotensin–aldosterone system, and is closely associated with chronic inflammation and increased oxidative stress [5,6,17–22]. In addition, tissue hypoxia and changes in blood flow patterns resulting from lower hemoglobin concentrations might also play an atherogenic role [5,6,17–22]. Our findings may have important clinical implications. Our results provide further evidence that anaemia may be an important and independent risk factor for all-cause and cardiovascular mortality. This makes anaemia another potential target for the treatment of type 2 diabetes. Two large randomized clinical trials of recombinant human erythropoietin in patients with stage 3 or 4 CKD – the cardiovascular risk reduction by early anaemia treatment with epoetin beta (CREATE) [23] and the correction of hemoglobin and outcomes in renal insufficiency (CHOIR) [24] – tested the hypothesis that the early and complete correction of anaemia with the use of recombinant human erythropoietin would result in improvements of cardiovascular outcomes, but both studies resulted in negative findings. In the CHOIR study [24], the largest study to date in CKD patients not receiving chronic dialysis, the use of a target hemoglobin level of 13.5 g/dl (as compared with 11.3 g/dl) was associated with increased risk of cardiovascular events and no incremental improvement in the quality of life. In the CREATE study [23], the early complete correction of anaemia did not significantly reduce the risk of future cardiovascular events during the 3 years of follow-up. A recent meta-analysis further confirmed an increased risk of all-cause mortality, arterio-venous access thrombosis and poorly controlled blood pressure in anaemic patients with CKD treated with erythropoietin [25]. The mechanisms that underlie this excess in cardiovascular events noted in the higher hemoglobin target group are unclear. Whether this increased risk relates to the achieved higher hemoglobin per se or to the means by which this was achieved – i.e., higher hemoglobin concentrations due to the use of higher doses of erythropoietin – remains unclear. Currently, there is no definitive evidence that the correction of anaemia in patients with type 2 diabetes and CKD leads to a reduction in the risk of cardiovascular and non-cardiovascular outcomes. Very recently, the results from the trial to reduce cardiovascular events with aranesp therapy (TREAT), the largest study to date in type 2 diabetic patients with stage 3 or 4 CKD and moderate anaemia who were not undergoing dialysis, showed that the use of darbepoetin alfa (to attempt to reach a target hemoglobin level of 13.0 g/dl) did not reduce the risk of death and major cardiovascular and renal events, and was associated with an increased risk of stroke [26]. However, it is also important to emphasize that there are no randomized controlled trials that have evaluated the effect of treatment of anaemia in type 2 diabetic patients without advanced CKD. Thus, one must be cautious to extrapolate the results from randomized controlled trials treating the severe degrees of anaemia associated with erythropoietin deficiency in advanced CKD to the less severe degrees of anaemia in patients with other causes of chronic anaemia. The major limitations of this study include a single baseline measurement of the variables of interest (including also eGFR), the lack of information on the specific causes of anaemia, a pos-
580
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
sible selection bias of excluding the patients who had missing hemoglobin measurement, an inability to adjust for certain specific anti-hypertensive agents (e.g., ACE-inhibitors, angiotensin receptor blockers or loop diuretics), and an issue of multiple testing with a propensity for false positive associations. Finally, we used an estimated GFR (i.e., the four-variable MDRD study equation) instead of a directly measured GFR to define CKD, and serum creatinine measurements were not calibrated to results obtained at the Cleveland Clinic Laboratory where serum creatinine was measured in the MDRD study. It is known that current GFR estimates have greater inaccuracy in populations without known CKD than in those with the disease [11]. Nonetheless, current GFR estimates facilitate the detection, evaluation, and management of CKD, and many organizations recommend the use of prediction equations for the evaluation of kidney function in large epidemiologic studies and in clinical practice [11,12]. Despite these limitations, our study has important strengths, including its prospective design, the large number of participants, and the ability to adjust for a wide range of established risk factors and potential confounders. Additionally, our patients were free of diagnosed cardiovascular disease, malignancy, end-stage renal disease and chronic obstructive pulmonary disease; the evaluation of patients with such complications would almost certainly have confounded interpretation of the data. Finally, we measured hemoglobin for diagnosing anaemia instead of hematocrit as previously made by other investigators [6,7]. Since hemoglobin is less sensitive than hematocrit to variations of cell volumes due to a variety of biological and pre-analytical variables, the measurement of hemoglobin is a more reliable parameter of anaemia than hematocrit [5]. In conclusion, our findings suggest that anaemia is associated with an increased risk of all-cause and cardiovascular mortality in a large cohort of type 2 diabetic individuals, independently of the presence of CKD and several other potential confounders. The advantage to treat anaemia in type 2 diabetes for reducing the risk of cardiovascular outcomes remains to be demonstrated. Conflict of interest None to declare. References [1] Al-Ahmad A, Rand WM, Manjunath G, et al. Reduced kidney function and anemia as risk factors for mortality in patients with left ventricular dysfunction. J Am Coll Cardiol 2001;38:955–62. [2] Jurkovitz CT, Abramson JL, Vaccarino LV, Weintraub WS, McClellan WM. Association of high serum creatinine and anemia increases the risk of coronary events: results from the prospective community-based atherosclerosis risk in communities (ARIC) study. J Am Soc Nephrol 2003;14:2919–25.
[3] Anand I, McMurray JJ, Whitmore J, et al. Anemia and its relationship to clinical outcome in heart failure. Circulation 2004;110:149–54. [4] Chonchol M, Nielson C. Hemoglobin levels and coronary artery disease. Am Heart J 2008;155:494–8. [5] KDOQI. KDOQI clinical practice guideline and clinical practice recommendations for anemia in chronic kidney disease: 2007 update of hemoglobin target. Am J Kidney Dis 2007;50:471–530. [6] Vlagopoulos PT, Tighiouart H, Weiner DE, et al. Anemia is a risk factor for cardiovascular disease and all-cause mortality in diabetes: the impact of chronic kidney disease. J Am Soc Nephrol 2005;16:3403–10. [7] Tong PC, Kong AP, So WY, et al. Hematocrit, independent of chronic kidney disease, predicts adverse cardiovascular outcomes in Chinese patients with type 2 diabetes. Diabetes Care 2006;29:2439–44. [8] Muggeo M, Verlato G, Bonora E, et al. The Verona Diabetes Study: a populationbased survey on known diabetes mellitus prevalence and 5-year all-cause mortality. Diabetologia 1995;38:318–25. [9] de Marco R, Locatelli F, Zoppini G, Verlato G, Bonora E, Muggeo M. Cause-specific mortality in type 2 diabetes. Diabetes Care 1999;22:756–61. [10] Targher G, Zoppini G, Chonchol M, et al. Glomerular filtration rate, albuminuria and risk of cardiovascular and all-cause mortality in type 2 diabetic individuals. Nutr Metab Cardiovasc Dis 2009; doi:10.1016/j.numecd.2009.10.002. [11] Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function-measured and estimated glomerular filtration rate. N Engl J Med 2006;354:2473–83. [12] American Diabetes Association. Standards of medical care in diabetes – 2009. Diabetes Care 2009;32(Suppl. 1):S13–61. [13] Robles NR, Angulo E, Grois J, Barquero A. Comparative effects of fosinopril and irbesartan on hematopoiesis in essential hypertensives. Ren Fail 2004;26:399–404. [14] Ting RZ, Szeto CC, Chan MH, Ma KK, Chow KM. Risk factors of vitamin B12 deficiency in patients receiving metformin. Arch Intern Med 2006;166:1975–9. [15] Macdougall IC. ACE-inhibitors and erythropoietin responsiveness. Am J Kidney Dis 2001;38:649–51. [16] Joss N, Patel R, Paterson K, Simpson K, Perry C, Stirling C. Anaemia is common and predicts mortality in diabetic nephropathy. QJM 2007;100:641–7. [17] Go AS, Yang J, Ackerson LM, et al. Hemoglobin level, chronic kidney disease, and the risks of death and hospitalization in adults with chronic heart failure: the anemia in chronic heart failure: outcomes and resource utilization (ANCHOR) study. Circulation 2006;113:2713–23. [18] Sarnak MJ, Tighiouart H, Manjunath G, et al. Anemia as a risk factor for cardiovascular disease in the atherosclerosis risk in communities (ARIC) study. J Am Coll Cardiol 2002;40:27–33. [19] Grune T, Sommerburg O, Siems WG. Oxidative stress in anemia. Clin Nephrol 2000;53(Suppl. 1):S18–22. [20] Eustace JA, Astor B, Muntner PM, Ikizler TA, Coresh J. Prevalence of acidosis and inflammation and their association with low serum albumin in chronic kidney disease. Kidney Int 2004;65:1031–40. [21] Chonchol M, Lippi G, Montagnana M, Muggeo M, Targher G. Association of inflammation with anaemia in patients with chronic kidney disease not requiring chronic dialysis. Nephrol Dial Transplant 2008;23:2879–83. [22] Mehdi U, Toto RD. Anemia, diabetes, and chronic kidney disease. Diabetes Care 2009;32:1320–6. [23] Drueke T, Locatelli F, Clyne N, et al. CREATE investigators. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. N Engl J Med 2006;355:2071–84. [24] Singh A, Saczech L, Tang KL, et al. CHOIR investigators. Correction of anemia with epoietin-alfa in chronic kidney disease. N Engl J Med 2006;355:2085–98. [25] Phrommintikul A, Haas SJ, Elsik M, Krum H. Mortality and target haemoglobin concentrations in anaemic patients with chronic kidney disease treated with erythropoietin: a meta-analysis. Lancet 2007;369:381–8. [26] Pfeffer MA, Burdmann EA, Chen CY, et al. TREAT investigators. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med 2009;361:2019–32.
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Anaemia, independent of chronic kidney disease, predicts all-cause and cardiovascular mortality in type 2 diabetic patients Giacomo Zoppini a,1 , Giovanni Targher a,∗,1 , Michel Chonchol b , Carlo Negri a , Vincenzo Stoico a , Isabella Pichiri a , Giuseppe Lippi c , Michele Muggeo a , Enzo Bonora a a Section of Endocrinology, Department of Biomedical and Surgical Sciences, University of Verona, Ospedale Civile Maggiore, Piazzale Stefani, 1, 37126 Verona, Italy b Division of Renal Diseases and Hypertension, University of Colorado Denver Health Sciences Center, Aurora, CO, USA c Section of Clinical Chemistry, Department of Biomedical and Morphological Sciences, University of Verona, Verona, Italy
a r t i c l e
i n f o
Article history: Received 22 September 2009 Received in revised form 2 December 2009 Accepted 2 December 2009 Available online 11 December 2009 Keywords: Anaemia Chronic kidney disease Type 2 diabetes Mortality Epidemiology
a b s t r a c t Objective: There is limited and controversial information on whether anaemia is a risk factor for cardiovascular mortality in type 2 diabetes, and whether this risk is modified by the presence of chronic kidney disease (CKD). We assessed the predictive role of lower hemoglobin concentrations on all-cause and cardiovascular mortality in a cohort of type 2 diabetic individuals. Methods: The cohort included 1153 type 2 diabetic outpatients, who were followed for a mean period of 4.9 years. The independent association of anaemia (i.e., hemoglobin <120 g/l in women and <130 g/l in men) with all-cause and cardiovascular mortality was evaluated by Cox proportional hazards regression models and adjusted for several potential confounders, including kidney function measures. Results: During follow-up, 166 (14.4%) patients died, 42.2% (n = 70) of them from cardiovascular causes. In univariate analysis, anaemia was associated with increased risk of all-cause (hazard ratio HR 2.62, 95% confidence intervals 1.90–3.60, p < 0.001) and cardiovascular mortality (HR 2.70, 1.67–4.37, p < 0.001). After adjustment for age, sex, body mass index, smoking, hypertension, dyslipidemia, diabetes duration, hemoglobin A1c, medication use (hypoglycemic, anti-hypertensive, lipid-lowering and anti-platelet drugs) and kidney function measures, the association of anaemia with all-cause (adjusted HR 2.11, 1.32–3.35, p = 0.002) and cardiovascular mortality (adjusted HR 2.23, 1.12–4.39, p = 0.020) remained statistically significant. Conclusions: Anaemia is associated with increased risk of all-cause and cardiovascular mortality in type 2 diabetic individuals, independently of the presence of CKD and other potential confounders. The advantage to treat anaemia in type 2 diabetes for reducing the risk of adverse cardiovascular outcomes remains to be demonstrated. © 2009 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Several epidemiologic studies have consistently shown that anaemia and chronic kidney disease (CKD) are both important risk factors of all-cause and cardiovascular mortality in the general population and in some groups of high-risk patients [1–5]. However, the information on the relationship among anaemia, CKD and cardiovascular outcomes in people with type 2 diabetes is limited and controversial, even though anaemia is common in type 2 diabetic patients with CKD. In fact, anaemia is often more severe and occurs at an earlier stage of CKD in patients with
∗ Corresponding author. Tel.: +39 045 8123748; fax: +39 045 8027314. E-mail address: [email protected] (G. Targher). 1 These authors contributed equally to this work. 0021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2009.12.008
type 2 diabetes compared with those with other causes of kidney disease [5]. In a recent study Vlagopoulos et al. [6] showed that anaemia, as defined by low levels of hematocrit (i.e., <36% in women and <39% in men), was a significant risk factor for cardiovascular disease and all-cause mortality primarily in diabetic patients who also had CKD (defined as an estimated glomerular filtration rate <60 ml/min/1.73 m2 ), whereas anaemia was not a risk factor for any outcome in those with normal or near-normal kidney function [6]. In contrast, Tong et al. reported that in a cohort of Chinese type 2 diabetic patients, who were followed for ∼3 years, the risk of future cardiovascular events increased progressively in presence of decreasing levels of hematocrit only in those without declining kidney function [7]. The aim of this study was to examine whether anaemia is associated with increased risk of all-cause and cardiovascular mortality
576
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
in a cohort of type 2 diabetes patients, and whether this risk is modified by the presence of CKD.
2. Methods The study was performed within the frame of the Verona Diabetes Study, an observational longitudinal study on chronic complications in type 2 diabetic outpatients attending the diabetes clinic at the University Hospital of Verona [8,9]. Data included in this analysis are based upon the cohort of Caucasian type 2 diabetic outpatients (n = 1153), who were recruited over the period of January 2000 to January 2002 and then followed-up until September 30, 2007, after excluding (i) patients who had a history of malignancy, chronic obstructive pulmonary disease, end-stage renal disease or cardiovascular disease (defined as angina, myocardial infarction, revascularization procedures and stroke), and (ii) those who had incomplete clinical and biochemical data for analysis. Type 2 diabetes was established when diagnosis was made after the age of 35 years, irrespective of treatment, or when the disease was treated with diet or oral hypoglycemic agents, irrespective of age at diagnosis. More details about the study design and recruitment methods have been reported elsewhere [10]. Briefly, the 1153 participants included in this analysis represented approximately 30% of the whole sample of type 2 diabetic outpatients (n = 3727), who regularly attended our diabetes clinic during years 2000–2002 and who had complete laboratory data for analysis after excluding patients with the above-mentioned comorbid conditions (n = 989, 26.5%). Notably, baseline demographics (∼50% male; mean age: 67 years vs. 66 years; mean body mass index: 27.9 kg/m2 vs. 28.2 kg/m2 ; mean diabetes duration: 16.0 years vs. 15.4 years and mean hemoglobin A1c 7.5% vs. 7.6%, respectively) as well as the overall crude rates of all-cause (14.4% vs. 14%) and cardiovascular (6.1% vs. 6.5%) mortality were not significantly different between the 1153 participants of the study and those with missing data for hemoglobin or other laboratory variables (n = 1585). All participants were periodically seen (every 3–6 months) for routine medical examinations of glycemic control and chronic complications of diabetes. The local ethics committee approved the study protocol. All participants gave their informed consent. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters. Blood pressure was measured with a standard mercury manometer. Information on comorbid conditions, current use of medications and smoking history was obtained in all patients by interviews during medical examinations. In all participants venous blood was withdrawn in the morning after an overnight fast for standard biochemical work-up. Hemoglobin concentration was determined by a photometrical technique on the fully automated hematological analyzer ADVIA 120-TM (Bayer Diagnostics, Milan, Italy). Serum creatinine (measured by using the Jaffé method – kinetic alkaline picrate assay), lipids and other biochemical blood measurements were determined by automatic colorimetric methods (DAX 96, Bayer Diagnostics, Milan, Italy). LDL-cholesterol was calculated by Friedewald’s equation, except in those with triglycerides exceeding 4.55 mmol/l (n = 36). Hemoglobin A1c was measured by a high-performance liquid chromatography analyzer (BioRad Diamat, Milan, Italy) and the upper limit of normality was 5.6%. Anaemia was defined as hemoglobin concentrations <120 g/l in women and <130 g/l in men, according to the World Health Organization criteria. Glomerular filtration rate was estimated (eGFR) from the four-variable Modification of Diet in Renal Disease (MDRD) equation [11] as follows: eGFR = 186 × (serum creatinine−1.154 ) × (age−0.203 ) × 1.212 (if black) × 0.742 (if female).
Urinary albumin excretion was measured from an early morning urine sample on at least three consecutive occasions, within a period of 4–6 months, as the albumin/creatinine ratio (ACR) by an immuno-nephelometric method. Microalbuminuria was defined as an ACR of 30–299 mg/g and macroalbuminuria as an ACR of ≥300 mg/g [12]. The presence of CKD was diagnosed when a patient had either eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria (i.e., micro- or macroalbuminuria) irrespective of eGFR values [11,12]. Hypertension was defined as a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg or current use of any treatment with anti-hypertensive medications. Information regarding specific classes of anti-hypertensive drugs (e.g., ACE-inhibitors and angiotensin receptor blockers) was not currently available. Atherogenic dyslipidemia was diagnosed when plasma triglycerides were ≥1.70 mmol/l and/or HDL-cholesterol was <1.04 mmol/l or when patients were taking lipid-lowering agents [12]. Vital status on September 30, 2007 was ascertained for all participants (n = 1153) by examining the database of the Social Health Unit of the Veneto Region, which includes all records of mortality occurred within the Veneto Region as well as the specific causes of death. Causes of death were identified in 100% of subjects. Trained nosologists coded death certificates using the International Classification of Diseases, Ninth Revision (ICD-9). Deaths were attributed to cardiovascular diseases when ICD-9 codes were 390–459. A selected sample of death certificates was reviewed manually to validate the process. 2.1. Statistical analysis Data are presented as means ± SD or proportions. Skewed variables were logarithmically transformed to improve normality prior to analyses. The unpaired t-test and the chi-squared test with Yates’s correction for continuity (for categorical variables) were used to compare the baseline clinical characteristics of participants stratified by the presence or absence of anaemia, defined as hemoglobin <120 g/l in women and <130 g/l in men. Univariate survival analysis stratified by anaemia and CKD status was performed by the Kaplan–Meier analysis, and the overall significance was calculated by the log-rank test. Cox regression analysis was used to study the effect of anaemia on the risk of all-cause and cardiovascular mortality after adjustment for potential confounders. Four forced entry multivariate regression models were performed. Hemoglobin was included as either continuous or dichotomous variable in these models. The first multivariate regression model was adjusted for age (years), gender (male vs. female) and BMI (kg/m2 ); the second model for age (years), gender (male vs. female), BMI (kg/m2 ), diabetes duration (years), hemoglobin A1c (%), hypertension (yes/no; see definition above), atherogenic dyslipidemia (yes/no; see definition above), LDL-cholesterol (mmol/l) and current use of medications (anti-platelet and hypoglycemic drugs) (yes/no); finally, the third and fourth regression models were further adjusted for eGFR and albuminuria, which were included as either categorical (i.e., eGFR <60 ml/min/1.73 m2 vs. ≥60 ml/min/1.73 m2 and ACR <30 mg/g vs. ≥30 mg/g, respectively) or continuous measures. The covariates included in multivariate regression models were chosen as potential confounders based on their biological plausibility or statistical association with mortality in univariate analysis. Results are presented as hazard ratios (HRs) with 95% confidence intervals (CI) and statistical significance was evaluated by the likelihood-ratio test. HRs for continuous variables were computed for each SD change. Statistical analyses were performed with statistical package SPSS 14.0. p-Values <0.05 were considered statistically significant.
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
577
Table 1 Baseline characteristics of type 2 diabetic patients stratified by anaemia status.
Sex (M/F) Age (years) Diabetes duration (years) Body mass index (kg/m2 ) Current smokers (%) Hypertension (%) Hemoglobin A1c (%) LDL-cholesterol (mmol/l) Triglycerides (mmol/l) HDL-cholesterol (mmol/l) Hemoglobin (g/l) eGFR (ml/min/1.73 m2 ) Micro- or macroalbuminuria (%) Chronic kidney disease (%)a Oral hypoglycaemic drugs (%) Insulin therapy (%) Anti-hypertensive drugs (%) Lipid-lowering drugs (%) Anti-platelet drugs (%)
No anaemia (n = 941)
Anaemia (n = 212)
p
477/464 67 ± 12 16 ± 9 28.1 ± 5.0 12.7 77 7.4 ± 1.5 3.41 ± 0.9 1.63 ± 1.1 1.41 ± 0.4 144.7 ± 12 71.6 ± 18 23.7 45.9 73.2 32.1 70.0 40.3 38.9
101/111 67 ± 15 15 ± 9 27.2 ± 4.5 29.8 82 7.6 ± 1.6 3.10 ± 0.9 1.71 ± 0.9 1.40 ± 0.4 106.1 ± 21 61.1 ± 24 32.2 67.9 65.6 50.5 73.1 33.5 42.5
0.45 0.52 0.01 0.03 0.001 0.12 0.06 <0.001 0.85 0.24 ND <0.001 0.03 <0.001 <0.001 <0.001 0.41 0.07 0.36
Cohort size, n = 1153. Anaemia was defined as hemoglobin <120 g/l in women and <130 g/l in men. Data are expressed as means ± SD or proportions. p-Values refer to t-test or the 2 test (for categorical variables). LDL, low density lipoprotein; HDL, high density lipoprotein; eGFR, estimated glomerular filtration rate; ND, not determined. a Chronic kidney disease was defined as eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria.
3. Results Overall, anaemia was present in 18.4% (n = 212; 52.4% female) of the whole cohort; mean [SD] hemoglobin: 106.1±21 g/l, median [interquartile range, IQR]: 113.5 [99–120] g/l in all anaemic participants; mean [SD]:101.2 ± 21 g/l, median [IQR]: 110.5 [87–116] g/l in anaemic women, and mean [SD]: 111.5 ± 21 g/l, median [IQR]: 119.0 [106–128] g/l in anaemic men. The baseline characteristics of participants stratified by anaemia status are shown in Table 1. Compared with those with normal hemoglobin levels, participants with anaemia had a greater frequency of CKD (defined as eGFR <60 ml/min/1.73 m2 and/or abnormal albuminuria), lower BMI, lower LDL-cholesterol, and tended to have higher hemoglobin A1c and greater prevalence of hypertension. In addition, they also had a shorter duration of diabetes and were more likely to be smokers and to be treated with insulin therapy compared with those without anaemia. Age, gender, triglycerides, HDL-cholesterol and proportions using lipidlowering, anti-platelet or anti-hypertensive medications were not significantly different between the groups. During the follow-up (mean ± SD: 4.9 ± 1.0, range: 1–6 years), 166 (14.4%) participants died, 70 (42.2%) of them from cardiovascular causes. The crude rates for all-cause and cardiovascular mortality were 58 (27.4%) and 26 (12.3%) in anaemic patients, and 108 (11.5%) and 44 (4.7%) in those without anaemia. Fig. 1 shows the results of the Kaplan–Meier survival analysis for all-cause mortality in participants stratified by anaemia and CKD status. Compared with those with no anaemia and no CKD (included as the referent group), participants with both anaemia and CKD had the highest risk of mortality, whereas those with anaemia or CKD alone had an intermediate risk (p < 0.001 by the log-rank test). Almost identical results were obtained for cardiovascular mortality (data not shown). The independent association of lower hemoglobin concentrations with all-cause and cardiovascular mortality was tested by Cox regression analysis (Table 2). In unadjusted regression models, lower hemoglobin was associated with a higher risk of all-cause and cardiovascular mortality when it was included as either continuous (for each SD decrement, i.e., 21 g/l) or dichotomous variable. Of note, the association of lower hemoglobin with all-cause and cardiovascular mortality remained significant even after adjustment for a broad spectrum of potential confounders, including the
presence of kidney dysfunction (Table 2, model 1–4). Others independent predictors of all-cause and cardiovascular mortality were older age, male sex, smoking history and presence of CKD (data not shown). We tested for a formal interaction between anaemia and eGFR <60 ml/min/1.73 m2 and between anaemia and abnormal albuminuria for each of the outcomes by including their crossproduct in the fully adjusted multivariate regression model. The anaemia × eGFR <60 ml/min/1.73 m2 interaction term (p = 0.084 and p = 0.395 for all-cause and cardiovascular mortality, respectively) and the anaemia × abnormal albuminuria interaction term (p = 0.385 and p = 0.174 for all-cause and cardiovascular mortality, respectively) were not statistically significant for any of the outcomes, except for a borderline significance for the anaemia–eGFR interaction term for all-cause mortality (p = 0.084). As shown in Table 3, we also examined the association of sexspecific quintiles of hemoglobin concentration with the risk of all-cause and cardiovascular mortality. The unadjusted HRs for allcause and cardiovascular mortality progressively increased with decreasing quintiles of hemoglobin concentration. After adjust-
Fig. 1. Kaplan–Meier survival analysis for all-cause mortality in 1153 type 2 diabetic patients stratified by anaemia and CKD status (no anaemia and no CKD as the reference category). The overall statistical significance was estimated by the log-rank test (p < 0.001).
578
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
Table 2 Cox proportional hazards models of lower hemoglobin levels for all-cause and cardiovascular mortality in type 2 diabetic patients. 1-SD decrement in hemoglobin
p
Anaemia (yes/no)
p
All-cause mortality Unadjusted HR (95% CI)
1.37 (1.20–1.53)
<0.001
2.62 (1.90–3.60)
<0.001
Adjusted HR (95% CI) Model 1 Model 2 Model 3 Model 4
1.45 (1.26–1.67) 1.41 (1.22–1.61) 1.24 (1.05–1.47) 1.18 (0.99–1.45)
<0.001 <0.001 0.010 0.07
2.39 (1.69–3.39) 2.18 (1.52–3.13) 2.11 (1.32–3.35) 1.91 (1.17–3.07)
<0.001 <0.001 0.002 0.009
Cardiovascular mortality Unadjusted HR (95% CI)
1.39 (1.18–1.67)
<0.001
2.70 (1.67–4.37)
<0.001
Adjusted HR (95% CI) Model 1 Model 2 Model 3 Model 4
1.45 (1.18–1.79) 1.37 (1.11–1.72) 1.35 (1.04–1.80) 1.30 (1.01–1.72)
0.001 0.004 0.032 0.050
2.31 (1.36–3.91) 2.10 (1.19–3.57) 2.23 (1.12–4.39) 1.89 (1.03–3.76)
0.002 0.009 0.020 0.043
Cohort size, n = 1153. Results are expressed as hazard ratios (HRs) and 95% confidence intervals (in parenthesis). Model 1: adjusted for age, sex and body mass index. Model 2: model 1 plus further adjustment for diabetes duration, hemoglobin A1c, smoking history, hypertension, atherogenic dyslipidemia, LDL-cholesterol, and current use of medications. Model 3: model 2 plus further adjustment for eGFR (<60 ml/min/1.73 m2 vs. >60 ml/min/1.73 m2 ) and albuminuria (ACR <30 mg/g vs. >30 mg/g) included as categorical variables. Model 4: model 2 plus further adjustment for eGFR and albuminuria included as continuous variables. In both the unadjusted regression models and the four multivariate regression models hemoglobin was included as either continuous (for each SD decrement) or dichotomous variable (i.e., hemoglobin <120 g/l in women and <130 g/l in men).
Table 3 Cox proportional hazards models of sex-specific quintiles of hemoglobin concentrations for all-cause and cardiovascular mortality in type 2 diabetic patients. I quintile <121 g/l W <132 g/l M
II quintile 121–131 g/l W 132–145 g/l M
III quintile 132–139 g/l W 146–151 g/l M
IV quintile 140–145 g/l W 152–157 g/l M
V quintile >146 g/l W >157 g/l M
All-cause mortality Unadjusted HR (95% CI) Fully adjusted HR (95% CI)
3.9 (2.3–6.5) 2.2 (1.3–3.9)
1.9 (1.1–3.4) 1.2 (0.7–2.1)
1.4 (0.9–2.5) 1.1 (0.5–1.9)
1.2 (0.8–2.4) 0.8 (0.4–1.7)
Ref. Ref.
<0.001 0.018
Cardiovascular mortality Unadjusted HR (95% CI) Fully adjusted HR (95% CI)
3.9 (1.8–8.2) 2.1 (1.2–5.3)
2.0 (1.1–4.4) 1.4 (0.8–3.6)
1.3 (0.8–3.2) 1.0 (0.6–2.7)
1.0 (0.5–2.5) 0.7 (0.4–1.5)
Ref. Ref.
<0.001 0.025
p for linear trend
Cohort size, n = 1153. Results are expressed as hazard ratios (HRs) and 95% confidence intervals (in parenthesis). W = women; M = men; Ref. = reference category. Fully adjusted regression models: adjustment for age, sex, body mass index, diabetes duration, hemoglobin A1c, smoking history, hypertension, atherogenic dyslipidemia, LDL-cholesterol, medication use, eGFR (<60 ml/min/1.73 m2 vs. >60 ml/min/1.73 m2 ) and albuminuria (ACR <30 mg/g vs. >30 mg/g).
ment for age, sex, BMI, smoking, hypertension, dyslipidemia, diabetes duration, hemoglobin A1c, medication use and kidney function measures, the risk of all-cause and cardiovascular mortality across hemoglobin quintiles was consistently attenuated and it was more pronounced among those belonging to the lowest hemoglobin quintile. Fig. 2 shows the age- and sex-adjusted association of anaemia with the risk all-cause mortality in subgroup analyses. Notably, anaemia was significantly associated with an increased risk of
Fig. 2. Subgroup analyses of the risk of all-cause mortality by anaemia. Hazard ratios (95% confidence intervals) are adjusted for age and sex. p-Values for the interaction term of anaemia with each stratification variable, which were observed in an ageand sex-adjusted regression model, are also reported.
mortality in all subgroups of participants, independently of their baseline values of BMI, hemoglobin A1c, eGFR and albuminuria. Almost identical results were observed for cardiovascular mortality (data not shown). 4. Discussion The major finding of this study is that anaemia, defined as hemoglobin <120 g/l in women and <130 g/l in men, is strongly associated with an increased risk of all-cause and cardiovascular mortality in a large outpatient cohort of type 2 diabetic individuals. The presence of anaemia roughly doubled the risk of dying, mainly from cardiovascular causes. Moreover, the risk of cardiovascular mortality increased approximately 40% for each SD decrement in hemoglobin concentration. Notably, the association of lower hemoglobin concentrations with the risk of all-cause and cardiovascular mortality was independent of a broad number of potential confounders, including traditional cardiovascular risk factors, diabetes-related variables, kidney function measures (eGFR and albuminuria) and use of medications, i.e., all factors potentially correlated also to raised hemoglobin concentrations [6,13–15]. Several prospective studies have shown that anaemia and CKD are both important risk factors of all-cause and cardiovascular mortality in the general population and in some groups of high-risk patients [1–5]. At present, however, the available information on the association among anaemia, CKD and adverse cardiovascular outcomes in people with type 2 diabetes is quantitatively limited and controversial.
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
In a small retrospective study of 508 type 2 diabetic patients with established nephropathy, it was found that anaemia strongly predicted all-cause mortality [16]. In a recent pooled analysis from four community-based studies of 3015 US diabetic individuals, who were followed for a median period of 8.6 years, anaemia was a risk factor for fatal and non-fatal cardiovascular events primarily in those who also had CKD. The combination of anaemia and CKD conferred a synergistic risk for adverse cardiovascular outcomes compared with each factor risk alone. In contrast, anaemia was not a significant risk factor for any outcome in those with normal or near-normal kidney function [6]. It is important to emphasize that the authors of this study diagnosed anaemia extrapolating hemoglobin concentrations from those of hematocrit, which were measured in four different laboratories, and did not adjust their results for some potentially important confounders, such as diabetes duration, hemoglobin A1c and medication use [6]. Moreover, the diagnosis of CKD was based only on eGFR values, whereas information on albuminuria was not available for diagnosing CKD as recommended by the recent National Kidney Foundation guidelines [11]. Conversely, in a cohort of 3983 Chinese type 2 diabetic patients who were followed for a median period of 3 years, lower levels of hematocrit were associated with increased risk of all-cause mortality and non-fatal cardiovascular events only in those without CKD [7]. In that study, however, the study cohort was hospital based, making selection bias a potential confounding factor. Moreover, the specific causes of death were not analysed, and the results were not adjusted for potentially important confounders, such as the use of hypoglycemic, lipid-lowering and anti-platelet drugs [7]. In the present study, we found that the association between anaemia and increased risk of all-cause and cardiovascular mortality was only partially modified by the presence of CKD. Although a large part of the excess hazards of hemoglobin for all-cause and cardiovascular mortality was likely explained by the presence of kidney dysfunction (from ∼20% to 50% when comparing model 4 to model 2 of Table 2), it appeared that anaemia is associated with increased risk of all-cause and cardiovascular mortality even in individuals with normal eGFR and normo-albuminuria but that this association may be stronger in those who also had CKD, as also confirmed in our subgroup analyses (Fig. 2). Interestingly, the anaemia–CKD interaction terms were not statistically significant for any of the study outcomes, except for a borderline significance for the anaemia–eGFR interaction term for all-cause mortality (p = 0.084). When we examined the association of sex-specific quintiles of hemoglobin concentration with mortality, the unadjusted HRs for all-cause and cardiovascular mortality increased progressively with decreasing quintiles of hemoglobin concentration. After adjustment for multiple potential confounders, including also kidney function measures, the risk of all-cause and cardiovascular mortality across hemoglobin quintiles was largely attenuated and it was more pronounced among those belonging to the lowest hemoglobin quintile (Table 3). Although the p-values for the linear trend in this analysis were statistically significant and suggested a possible linearity of the hemoglobin concentration with mortality risks, a potential threshold below the IIIrd quintile of hemoglobin cannot be definitely excluded from these data. Anaemia might be just a marker of underlying established risk factors or important comorbidities. Notably, in the present study we excluded patients who had a history of malignancy, severe chronic obstructive pulmonary disease, end-stage renal disease or cardiovascular disease. Moreover, we found that lower hemoglobin concentration was associated with increased risk of all-cause and cardiovascular mortality independently of several established risk factors and potential confounders. Thus, it is conceivable that anaemia might confer an excess risk over and above the risk attributable to underlying established risk factors. The recognition of the biological mechanism(s) linking anaemia
579
to cardiovascular mortality is beyond the scope of this study. However, some putative biological mechanisms might be mentioned. It is known that anaemia may result in many potentially adverse effects on the cardiovascular system, such as increased cardiac output and increased heart rate. Over the long term, adaptations that initially increase cardiac output lead to left ventricular enlargement and eccentric left ventricular hypertrophy. Thus, it is likely that anaemic patients with left ventricular hypertrophy or subclinical ischemic heart disease might be at higher risk of developing symptomatic cardiac disease [4–6,17–22]. Anaemia may activate the sympathetic nervous system, stimulate the renin–angiotensin–aldosterone system, and is closely associated with chronic inflammation and increased oxidative stress [5,6,17–22]. In addition, tissue hypoxia and changes in blood flow patterns resulting from lower hemoglobin concentrations might also play an atherogenic role [5,6,17–22]. Our findings may have important clinical implications. Our results provide further evidence that anaemia may be an important and independent risk factor for all-cause and cardiovascular mortality. This makes anaemia another potential target for the treatment of type 2 diabetes. Two large randomized clinical trials of recombinant human erythropoietin in patients with stage 3 or 4 CKD – the cardiovascular risk reduction by early anaemia treatment with epoetin beta (CREATE) [23] and the correction of hemoglobin and outcomes in renal insufficiency (CHOIR) [24] – tested the hypothesis that the early and complete correction of anaemia with the use of recombinant human erythropoietin would result in improvements of cardiovascular outcomes, but both studies resulted in negative findings. In the CHOIR study [24], the largest study to date in CKD patients not receiving chronic dialysis, the use of a target hemoglobin level of 13.5 g/dl (as compared with 11.3 g/dl) was associated with increased risk of cardiovascular events and no incremental improvement in the quality of life. In the CREATE study [23], the early complete correction of anaemia did not significantly reduce the risk of future cardiovascular events during the 3 years of follow-up. A recent meta-analysis further confirmed an increased risk of all-cause mortality, arterio-venous access thrombosis and poorly controlled blood pressure in anaemic patients with CKD treated with erythropoietin [25]. The mechanisms that underlie this excess in cardiovascular events noted in the higher hemoglobin target group are unclear. Whether this increased risk relates to the achieved higher hemoglobin per se or to the means by which this was achieved – i.e., higher hemoglobin concentrations due to the use of higher doses of erythropoietin – remains unclear. Currently, there is no definitive evidence that the correction of anaemia in patients with type 2 diabetes and CKD leads to a reduction in the risk of cardiovascular and non-cardiovascular outcomes. Very recently, the results from the trial to reduce cardiovascular events with aranesp therapy (TREAT), the largest study to date in type 2 diabetic patients with stage 3 or 4 CKD and moderate anaemia who were not undergoing dialysis, showed that the use of darbepoetin alfa (to attempt to reach a target hemoglobin level of 13.0 g/dl) did not reduce the risk of death and major cardiovascular and renal events, and was associated with an increased risk of stroke [26]. However, it is also important to emphasize that there are no randomized controlled trials that have evaluated the effect of treatment of anaemia in type 2 diabetic patients without advanced CKD. Thus, one must be cautious to extrapolate the results from randomized controlled trials treating the severe degrees of anaemia associated with erythropoietin deficiency in advanced CKD to the less severe degrees of anaemia in patients with other causes of chronic anaemia. The major limitations of this study include a single baseline measurement of the variables of interest (including also eGFR), the lack of information on the specific causes of anaemia, a pos-
580
G. Zoppini et al. / Atherosclerosis 210 (2010) 575–580
sible selection bias of excluding the patients who had missing hemoglobin measurement, an inability to adjust for certain specific anti-hypertensive agents (e.g., ACE-inhibitors, angiotensin receptor blockers or loop diuretics), and an issue of multiple testing with a propensity for false positive associations. Finally, we used an estimated GFR (i.e., the four-variable MDRD study equation) instead of a directly measured GFR to define CKD, and serum creatinine measurements were not calibrated to results obtained at the Cleveland Clinic Laboratory where serum creatinine was measured in the MDRD study. It is known that current GFR estimates have greater inaccuracy in populations without known CKD than in those with the disease [11]. Nonetheless, current GFR estimates facilitate the detection, evaluation, and management of CKD, and many organizations recommend the use of prediction equations for the evaluation of kidney function in large epidemiologic studies and in clinical practice [11,12]. Despite these limitations, our study has important strengths, including its prospective design, the large number of participants, and the ability to adjust for a wide range of established risk factors and potential confounders. Additionally, our patients were free of diagnosed cardiovascular disease, malignancy, end-stage renal disease and chronic obstructive pulmonary disease; the evaluation of patients with such complications would almost certainly have confounded interpretation of the data. Finally, we measured hemoglobin for diagnosing anaemia instead of hematocrit as previously made by other investigators [6,7]. Since hemoglobin is less sensitive than hematocrit to variations of cell volumes due to a variety of biological and pre-analytical variables, the measurement of hemoglobin is a more reliable parameter of anaemia than hematocrit [5]. In conclusion, our findings suggest that anaemia is associated with an increased risk of all-cause and cardiovascular mortality in a large cohort of type 2 diabetic individuals, independently of the presence of CKD and several other potential confounders. The advantage to treat anaemia in type 2 diabetes for reducing the risk of cardiovascular outcomes remains to be demonstrated. Conflict of interest None to declare. References [1] Al-Ahmad A, Rand WM, Manjunath G, et al. Reduced kidney function and anemia as risk factors for mortality in patients with left ventricular dysfunction. J Am Coll Cardiol 2001;38:955–62. [2] Jurkovitz CT, Abramson JL, Vaccarino LV, Weintraub WS, McClellan WM. Association of high serum creatinine and anemia increases the risk of coronary events: results from the prospective community-based atherosclerosis risk in communities (ARIC) study. J Am Soc Nephrol 2003;14:2919–25.
[3] Anand I, McMurray JJ, Whitmore J, et al. Anemia and its relationship to clinical outcome in heart failure. Circulation 2004;110:149–54. [4] Chonchol M, Nielson C. Hemoglobin levels and coronary artery disease. Am Heart J 2008;155:494–8. [5] KDOQI. KDOQI clinical practice guideline and clinical practice recommendations for anemia in chronic kidney disease: 2007 update of hemoglobin target. Am J Kidney Dis 2007;50:471–530. [6] Vlagopoulos PT, Tighiouart H, Weiner DE, et al. Anemia is a risk factor for cardiovascular disease and all-cause mortality in diabetes: the impact of chronic kidney disease. J Am Soc Nephrol 2005;16:3403–10. [7] Tong PC, Kong AP, So WY, et al. Hematocrit, independent of chronic kidney disease, predicts adverse cardiovascular outcomes in Chinese patients with type 2 diabetes. Diabetes Care 2006;29:2439–44. [8] Muggeo M, Verlato G, Bonora E, et al. The Verona Diabetes Study: a populationbased survey on known diabetes mellitus prevalence and 5-year all-cause mortality. Diabetologia 1995;38:318–25. [9] de Marco R, Locatelli F, Zoppini G, Verlato G, Bonora E, Muggeo M. Cause-specific mortality in type 2 diabetes. Diabetes Care 1999;22:756–61. [10] Targher G, Zoppini G, Chonchol M, et al. Glomerular filtration rate, albuminuria and risk of cardiovascular and all-cause mortality in type 2 diabetic individuals. Nutr Metab Cardiovasc Dis 2009; doi:10.1016/j.numecd.2009.10.002. [11] Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function-measured and estimated glomerular filtration rate. N Engl J Med 2006;354:2473–83. [12] American Diabetes Association. Standards of medical care in diabetes – 2009. Diabetes Care 2009;32(Suppl. 1):S13–61. [13] Robles NR, Angulo E, Grois J, Barquero A. Comparative effects of fosinopril and irbesartan on hematopoiesis in essential hypertensives. Ren Fail 2004;26:399–404. [14] Ting RZ, Szeto CC, Chan MH, Ma KK, Chow KM. Risk factors of vitamin B12 deficiency in patients receiving metformin. Arch Intern Med 2006;166:1975–9. [15] Macdougall IC. ACE-inhibitors and erythropoietin responsiveness. Am J Kidney Dis 2001;38:649–51. [16] Joss N, Patel R, Paterson K, Simpson K, Perry C, Stirling C. Anaemia is common and predicts mortality in diabetic nephropathy. QJM 2007;100:641–7. [17] Go AS, Yang J, Ackerson LM, et al. Hemoglobin level, chronic kidney disease, and the risks of death and hospitalization in adults with chronic heart failure: the anemia in chronic heart failure: outcomes and resource utilization (ANCHOR) study. Circulation 2006;113:2713–23. [18] Sarnak MJ, Tighiouart H, Manjunath G, et al. Anemia as a risk factor for cardiovascular disease in the atherosclerosis risk in communities (ARIC) study. J Am Coll Cardiol 2002;40:27–33. [19] Grune T, Sommerburg O, Siems WG. Oxidative stress in anemia. Clin Nephrol 2000;53(Suppl. 1):S18–22. [20] Eustace JA, Astor B, Muntner PM, Ikizler TA, Coresh J. Prevalence of acidosis and inflammation and their association with low serum albumin in chronic kidney disease. Kidney Int 2004;65:1031–40. [21] Chonchol M, Lippi G, Montagnana M, Muggeo M, Targher G. Association of inflammation with anaemia in patients with chronic kidney disease not requiring chronic dialysis. Nephrol Dial Transplant 2008;23:2879–83. [22] Mehdi U, Toto RD. Anemia, diabetes, and chronic kidney disease. Diabetes Care 2009;32:1320–6. [23] Drueke T, Locatelli F, Clyne N, et al. CREATE investigators. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. N Engl J Med 2006;355:2071–84. [24] Singh A, Saczech L, Tang KL, et al. CHOIR investigators. Correction of anemia with epoietin-alfa in chronic kidney disease. N Engl J Med 2006;355:2085–98. [25] Phrommintikul A, Haas SJ, Elsik M, Krum H. Mortality and target haemoglobin concentrations in anaemic patients with chronic kidney disease treated with erythropoietin: a meta-analysis. Lancet 2007;369:381–8. [26] Pfeffer MA, Burdmann EA, Chen CY, et al. TREAT investigators. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med 2009;361:2019–32.