SEARCH

SEARCH BY CITATION

Keywords:

  • biomarkers;
  • cardiovascular system;
  • galectin-3;
  • general population;
  • prognosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

Abstract.  de Boer RA, van Veldhuisen DJ, Gansevoort RT, Muller Kobold AC, van Gilst WH, Hillege HL, Bakker SJL, van der Harst P (University of Groningen). The fibrosis marker galectin-3 and outcome in the general population. J Intern Med 2012; 272: 55–64.

Objective.  Galectin-3 is involved in fibrosis and inflammation and plays a role in heart failure, renal disease, obesity and cancer. We aimed to establish the relationship between galectin-3 and cardiovascular (CV) risk factors and mortality in the general population.

Design and subjects.  This study included 7968 subjects from the Prevention of REnal and Vascular ENd-stage Disease (PREVEND) cohort, with a median follow-up of approximately 10 years. Plasma galectin-3 was measured in baseline samples.

Main outcome measures.  We investigated the relationships between galectin-3 levels, demographic characteristics and risk factors of CV disease. We determined the prognostic value for all-cause, CV and cancer mortality.

Results.  The mean age of the population was 50 ± 13 years. Mean blood pressure was 129/74 mmHg, mean cholesterol was 5.7 ± 1.1 mmol L−1 and median galectin-3 was 10.9 ng mL−1 [interquartile range (IQR) 9.0–13.1]. Galectin-3 levels correlated with a wide range of risk factors of CV disease, including blood pressure, serum lipids, body mass index, renal function and N-terminal pro-B-type natriuretic peptide (< 0.0001). We observed a strong association between galectin-3 and age. Furthermore, we found a gender interaction, with female subjects (= 4001) having higher median galectin-3 levels (11.0 ng mL−1, IQR 9.1–13.4 vs. men (= 3967) 10.7 ng mL−1, IQR 8.9–12.8; < 0.0001), and galectin-3 levels in women more strongly correlated with risk factors of CV disease. After correction for the classical CV risk factors (smoking, blood pressure, cholesterol and diabetes), galectin-3 levels independently predicted all-cause mortality (hazard ratio per SD galectin-3 1.09, 95% CI 1.01–1.19; = 0.036), but not CV and cancer mortality separately.

Conclusions.  Galectin-3 is associated with age and risk factors of CV disease, with a strong gender interaction for these correlations. Galectin-3 predicts all-cause mortality in the general population.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

Galectin-3 (LGALS3) is a member of the galectin family, which comprises β-galactoside lectins. Galectin-3 is mainly produced and secreted by activated macrophages, mast cells and eosinophils, but also by other tissues such as the epithelium of the gastrointestinal and respiratory tracts. Galectin-3 is ligand-activated by binding to oligosaccharides via its carbohydrate-recognition domain, which has strong affinity for β-galactosides. Several ligands for galectin-3 have been identified, including various glycosylated matrix proteins, such as laminin, fibronectin and integrins [1, 2]. The cellular actions of galectin-3 result in cell adhesion and proliferation, which may translate into physiological processes such as fibrosis and cancer progression. Indeed, galectin-3 has been shown to play a central role in fibrosis and tissue remodelling [1, 2]. In addition, galectin-3 plays a systemic role in inflammatory and proliferation responses [1–4]. In line with this, galectins have been implicated in various diseases, particularly in cancer [5, 6], immunological disorders [3, 4, 7] and common multifactor cardiovascular (CV) disease such as obesity [8] and heart failure (HF) [9].

The level of soluble galectin-3 can be measured in peripheral blood and is emerging as a novel biomarker of fibrosis. In HF, plasma galectin-3 is a strong prognostic marker of outcome [10–12]. Furthermore, circulating galectin-3 is substantially elevated in patients with cancer, particularly metastatic cancer [13], and it has been suggested that galectin-3 levels could guide therapy [14].

Previous studies have focused on populations with a specific condition. We herein report for the first time the relation between plasma galectin-3 and established risk factors for CV disease and the predictive value of this lectin for mortality in a large population-based sample of almost 8,000 subjects.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

PREVEND

The Prevention of REnal and Vascular ENd-stage Disease (PREVEND) study has been designed to prospectively investigate the natural course of increased levels of urinary albumin excretion (UAE) and the relation to renal and CV disease in a large cohort drawn from the general population. The PREVEND study has been described in detail elsewhere [15, 16]. During the period 1997–1998, 8592 subjects were enrolled in the PREVEND cohort. Study subjects were asked to attend the outpatient clinic between 8:00 a.m. and 1:00 p.m., to fast overnight (no eating or drinking) prior to their visit and to bring a urine sample. Blood was taken, anticoagulated with EDTA and stored at −80 °C until analysis. A plasma sample was available for 8345 subjects. We excluded non-Caucasian patients (= 369) because they differed considerably with respect to CV disease risk profile, including serum lipids, blood pressure and blood glucose. A few samples could not be evaluated (= 8); thus, 7968 subjects were available for analysis. The PREVEND study was approved by the local medical ethics committee and was conducted in accordance with the Declaration of Helsinki. All subjects provided written informed consent.

Measurement of galectin-3

Galectin-3 was measured using an enzyme-linked immunosorbent assay (ELISA; BG Medicine, Inc., Waltham, MA, USA). This assay quantitatively measures the concentration of human galectin-3 levels in EDTA-anticoagulated plasma. The assay has a high sensitivity (lower detection limit of 1.13 ng mL−1) and exhibits no cross-reactivity with collagens or other members of the galectin family [17]. Commonly used medications such as angiotensin-converting enzyme inhibitors, β-blockers, spironolactone, furosemide, aspirin, warfarin, coumarines and digoxin do not interfere with the assay [17]. All samples were assayed in duplicate. Two standard controls were included in all runs: a lower control (expected value 13.0–23.1 ng mL−1) and an upper control (expected value 48.9–81.5 ng mL−1). The average lower control values were 16.65 ± 1.13 ng mL−1 (coefficient of variance 6.8%), and the average upper control values were 68.17 ± 3.20 ng mL−1 (coefficient of variance 4.7%).

Definitions

Ten blood pressure measurements were recorded during 10 min (with an automatic Dinamap XL Model 9300 series device); systolic and diastolic blood pressures were calculated as the mean of the last two measurements. Hypertension was defined as having a systolic blood pressure >140 mmHg and/or a diastolic blood pressure >90 mmHg and/or use of antihypertensive medication. Body mass index (BMI) was calculated as weight divided by height squared (kg m−2). Type 2 diabetes was defined as a fasting glucose level of >7.0 mmol L−1 (126 mg dL−1) or a nonfasting glucose level of >11.1 mmol L−1 (200 mg dL−1) or the use of antidiabetic drugs. Hypercholesterolaemia was defined as a serum cholesterol level of >6.5 mmol L−1 (251 mg dL−1) or, in the presence of a history of myocardial infarction (MI), serum cholesterol >5.0 mmol L−1 (193 mg dL−1), or the use of lipid-lowering medication. Smoking was defined as current smoking or smoking within the previous year. History of MI or cerebrovascular disease was defined as participant-reported hospitalization for at least 3 days as a result of these conditions. A C-reactive protein level >3 mg L−1 was defined as elevated. N-terminal pro-B-type natriuretic peptide (NT-proBNP) (pg mL−1) was measured as previously described [16]. UAE was calculated as the average value from two consecutive 24-h urine collections. Glomerular filtration rate was estimated (eGFR) using the simplified Modification of Diet in Renal Disease (sMDRD) formula [18].

Mortality

Mortality data were collected through the municipal register. Cause of death was obtained by linking the death certificate number to the primary cause of death as coded by the Dutch Central Bureau of Statistics. Survival time was defined as the period from the date of urine collection to the date of incident mortality, until 1 January 2009. If a person moved to an unknown destination, the date on which that individual could no longer be tracked via the municipal registry was used as the censor date.

Statistical analyses

Normally distributed variables are expressed as means ± standard deviation (SD). Non-normally distributed variables are presented as medians [interquartile range (IQR)] and were log transformed before regression analyses. Galectin-3 was first log-transformed and then normalized to mean = 0 and SD = 1 to facilitate interpretation. Clinical and biochemical characteristics were compared across quintiles of galectin-3 levels with analysis of variance (anova) for continuous variables and chi-square test for categorical variables. A multivariable regression model was developed using bootstrap selection methods with all baseline characteristics (Table 1) except the MDRD formula which is an estimate based on other variables (age, gender and creatinine level). We selected the covariates for our full model based on 1000 bootstrap samples using a backward stepwise algorithm. The standard errors of the full model were also estimated by 1000 bootstrap replicates.

Table 1.   Baseline characteristics
Median (IQR) galectin-3 (ng mL−1) CharacteristicTotal (= 7968)Quintiles of galectin-3P
12345
10.9 (9.0–13.1)7.7 (7.0–8.2)9.4 (9.0–9.8)10.9 (10.5–11.3)12.6 (12.2–13.1)15.6 (14.5–17.7)
  1. DM, diabetes mellitus; HC, hypercholesterolaemia; WHR, waist–hip ratio; BMI, body mass index (kg m−2); BP, blood pressure (mm Hg); cholesterol, total cholesterol (mmol L−1); LDL, low-density lipoprotein cholesterol (mmol L−1); HDL, high-density lipoprotein cholesterol (mmol L−1); triglycerides (mmol L−1, fasting); glucose (mmol L−1, fasting); creatinine, serum creatinine (μmol L−1); cystatin-C (mg L−1); MDRD, estimated glomerular filtration rate (eGFR), according to the Modification of Diet in Renal Disease formula (mL min−1); UAE, urinary albumin excretion rate (mg per 24 h); hs-CRP, high-sensitivity C-reactive protein (mg L−1); NT-proBNP, N-terminal pro-B-type natriuretic peptide (pg mL−1).

Age (years)49.5 ± 12.744.4 ± 11.145.9 ± 11.348.8 ± 12.052.1 ± 12.656.4 ± 12.80.000
Female gender (%)50.247.448.948.350.656.10.000
Smoking (last 5 years, %)44.647.446.344.344.440.60.002
DM (%)3.62.32.33.14.36.10.000
MI (%)3.71.82.42.74.27.40.000
Hypertension (%)33.422.226.631.139.747.90.000
HC (%)34.324.829.734.738.643.90.000
Stroke (%)0.90.80.60.61.31.60.004
WHR0.88 ± 0.100.87 ± 0.100.87 ± 0.100.88 ± 0.100.89 ± 0.090.90 ± 0.090.000
BMI26.1 ± 4.2025.1 ± 3.825.5 ± 2.926.2 ± 4.126.6 ± 4.427.0 ± 4.40.000
Systolic BP129.2 ± 20.2125.0 ± 18.1126.6 ± 19.0128.6 ± 19.6131.3 ± 20.6134.9 ± 22.50.000
Diastolic BP74.0 ± 9.772.1 ± 9.473.3 ± 9.874.1 ± 9.675.2 ± 9.875.4 ± 9.80.000
Cholesterol5.66 ± 1.125.41 ± 1.055.56 ± 1.105.68 ± 1.115.73 ± 1.115.91 ± 1.170.000
LDL3.69 ± 1.053.47 ± 1.003.60 ± 1.013.71 ± 1.043.77 ± 1.053.90 ± 1.060.000
HDL1.27 [1.03–1.56]1.32 [1.07–1.62]1.28 [1.04–1.57]1.25 [1.03–1.55]1.24 [1.03–1.53]1.24 [0.99–1.52]0.000
Triglycerides1.16 [0.85–1.68]1.02 [0.75–1.43]1.11 [0.82–1.59]1.17 [0.86–1.69]1.23 [0.89–1.78]1.31 [0.95–1.92]0.000
Glucose4.7 [4.4–5.1]4.6 [4.3–5.0]4.7 [4.3–5.1]4.7 [4.4–5.1]4.8 [4.4–5.2]4.9 [4.5–5.3]0.000
Creatinine82 [74–92]80 [73–89]81 [72–90]83 [74–92]84 [74–93]84 [75–97]0.000
Cystatin-C0.78 [0.69–0.88]0.73 [0.65–0.82]0.75 [0.67–0.83]0.77 [0.69–0.87]0.80 [0.72–0.90]0.86 [0.74–1.00]0.000
MDRD80 [71–89]84 [76–93]83 [75–91]80 [72–89]78 [69–86]73 [64–83]0.000
UAE9.5 [6.4–17.8]9.1 [6.4–15.3]8.9 [6.2–14.9]9.1 [6.3–17.2]10.2 [6.5–19.7]10.4 [6.4–24.2]0.000
hs-CRP1.29 [0.56–3.00]0.89 [0.39–2.16]1.04 [0.49–2.40]1.33 [0.58–2.92]1.53 [0.71–3.42]1.98 [0.85–4.28]0.000
NT-proBNP38.6 [17.4–74.9]32.8 [15.4–61.4]32.1 [14.7–62.2]37.2 [16.4–71.1]42.5 [19.2–82.8]52.7 [25.2–108.0]0.000

Time to event between quintiles of galectin-3 was compared with Kaplan–Meier curves and log-rank test. The Cox proportional hazards model was used to calculate the hazard ratio and 95% confidence intervals (CIs). Sequential models were fitted with the first model including no covariates (unadjusted), the second model adjusted for age and gender and the third model adjusted for age, gender, previous MI, previous stroke, hypertension, hypercholesterolaemia and diabetes. The assumptions underlying the proportional hazards model were tested and found to be valid.

All reported probability values are two-tailed, and < 0.05 was considered statistically significant. All analyses were performed using stata version 11.0 for Windows software (StataCorp, College Station, TX, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

Study population and baseline characteristics

The median galectin-3 level was 10.9 (IQR 9.0–13.1) ng mL−1. Baseline clinical characteristics and biochemistry of all subjects according to quintiles of galectin-3 levels are presented in Table 1. All demographic and clinical characteristics were statistically different across galectin-3 quintiles (< 0.05, anova and chi-square tests for continuous and categorical variables, respectively). Overall, the subjects in the higher quintiles were more likely to be older and female, and they had a higher burden of CV diseases and risk factors, including elevated blood pressure, dyslipidaemia, kidney dysfunction and elevated high-sensitivity C-reactive protein (hs-CRP) and NT-proBNP levels (Table 1; gender-stratified data are shown in Table S1).

Association between galectin-3 levels and clinical characteristics

Galectin-3 levels were significantly associated with multiple clinical and biochemical baseline characteristics. Age explained 9.9% of the variance of galectin-3 levels (Table 2). However, the variance of the galectin-3 levels explained by other baseline characteristics was small; renal function parameters had the strongest β coefficients (Table 2). We explored whether the inclusion of a squared or cubed age term added to the model, and we found that a squared age term added significantly (< 0.001), whereas the cubed term did not. This suggests a log-squared relationship between galectin-3 levels and age (Fig. 1 and Figure S1). In our basic model, adjusted for age, age-squared and gender, we tested whether clinical and biochemical baseline characteristics remained independently associated with galectin-3 levels (Table 2). Next, we performed a multivariable analysis using a bootstrap selection method and identified gender, age-squared, BMI, triglycerides, creatinine, cystatin-C and hs-CRP as independent correlates of galectin-3 levels appearing in more than 70% of the bootstrap samples (Table 3 and Table S2). The full model explained 16.3% of the variance of galectin-3 levels in our population. Subjects using antihypertensive medication, lipid-lowering drugs and oral antidiabetic agents had higher baseline galectin-3 levels (< 0.0001 compared with subjects not using medication); this may be associated with the risk factors (blood pressure, hypercholesterolaemia and diabetes) for which these drugs were prescribed (data not shown).

Table 2.   Relationship between galectin-3 levels and baseline characteristics
CharacteristicUnivariable Std (beta)R2P-valueAdjusted* Std (beta)R2P-value
  1. Abbreviations: see Table 1. *Adjusted for age, age-squared and gender. Standardized β coefficient (Std-beta) reflects the change in the dependent variable for 1-SD change in the independent variable. A high Std-beta value reflects greater strength of the association.

Age0.3150.100.000   
Female gender0.0480.0020.000   
Smoking (last 5 years)−0.0510.0030.0000.0020.1090.876
DM0.0670.0050.0000.0110.1090.302
MI0.0950.0090.0000.0280.1100.009
Hypertension0.1810.0330.0000.0410.1100.001
HC0.1320.0170.0000.0480.1110.000
Stroke0.0400.0020.0000.0200.1090.062
WHR0.1100.0120.0000.0880.1130.000
BMI0.1630.0260.0000.0930.1170.000
Systolic BP0.1620.0260.0000.0310.1100.013
Diastolic BP0.1100.0120.0000.0180.1090.128
Cholesterol0.1380.0190.0000.0490.1110.000
LDL0.1310.0170.0000.0490.1120.000
HDL−0.0730.0050.000−0.0970.1170.000
Triglycerides0.1090.0120.0000.0900.1180.000
Glucose0.0990.0100.0000.0230.1100.051
Creatinine0.1820.0330.0000.1920.1390.000
Cystatin-C0.2640.0700.0000.1900.1420.000
MDRD−0.2760.0760.000−0.1650.1330.000
UAE0.0750.0060.0000.0460.1110.000
CRP0.1250.0160.0000.0830.1160.000
NT-proBNP0.0970.0090.0000.0760.1160.000
image

Figure 1.  Graph showing galectin-3 levels in male (blue line) and female subjects (red line). Grey-shaded areas indicate 95% confidence intervals. Galectin-3 levels increase with increasing age, particularly in female subjects.

Download figure to PowerPoint

Table 3.   Multivariable model
VariableBeta ± SEStd (beta)P-value
  1. Standardized β coefficient (Std-beta) reflects the change in the dependent variable for 1-SD change in the independent variable. A large Std-beta reflects greater strength of the association.

Age20.00019 ± 8.96e-060.2433.90e-94
Gender0.317 ± 0.0250.1598.17e-41
BMI0.014 ± 0.0030.0593.91e-07
Triglycerides0.058 ± 0.0120.0573.60e-07
Creatinine0.007 ± 0.00070.1311.67e-21
Cystatin-C0.533 ± 0.0640.1126.11e-17
hs-CRP0.010 ± 0.0020.0531.08e-06

Gender interaction

Next, we explored the possibility of a gender interaction. We observed a strong gender * age interaction with galectin-3 levels (= 2.39e-04; Fig. 1). We also observed interactions of gender with hypertension (= 0.021), BMI (= 5.08e-05), systolic blood pressure (= 4.56e-03), LDL cholesterol (= 0.027) and triglycerides (= 4.19e-05) (Table S3 and Figure S2). We tested the interaction terms of gender with age, triglycerides and BMI combined in our full multivariable model. The gender * triglyceride interaction term lost significance, but the gender * age and gender * BMI terms both remained significant (= 0.027 and = 8.63e-05, respectively).

Mortality

Among 7968 subjects, a total of 613 (7.7%) died: 181 due to CV causes (2.3%), 289 due to cancer (3.6%) and 143 (1.8%) due to other causes during 77 040 person-years at risk.

Mortality is shown by the Kaplan–Meier survival curves in Fig. 2a–c for galectin-3 quintiles. All-cause, CV and cancer mortality were increased, especially in the highest quintile of galectin-3 (Fig. 2). The hazard ratio (HR) for all-cause mortality was 1.46 per SD increase in galectin-3 (95% CI 1.37–1.56; < 0.0001; Table 4). Cox regression models were adjusted for the classical CV disease risk factors. After correction for age, gender, hypertension, hypercholesterolaemia and diabetes, one SD increase in galectin-3 was associated with a 9% increased risk of all-cause mortality (HR 1.09, 95% CI 1.01–1.19; = 0.036). Crude HRs for CV and cancer mortality were 1.56 (95% CI 1.39–1.75) and 1.41 (95% CI 1.28–1.56), respectively; after correction, the corresponding values were 1.10 (95% CI 0.94–1.28) and 1.08 (95% CI 0.96–1.22). Galectin-3 no longer predicted outcome after adjustment for other markers of CV mortality, such as UAE and NT-proBNP (Table 4).

image

Figure 2.  (a–c) Kaplan–Meier curves showing all-cause mortality (a), cardiovascular mortality (b) and cancer mortality (c) in study subjects according to quintiles of galectin-3. Subjects showed increased mortality with increasing galectin-3 levels; this relation was strongest for all-cause mortality (χ2 = 167.78; < 0.0001).

Download figure to PowerPoint

Table 4.   Galectin-3 and risk of mortality
ModelAll cause HR (95% CI)P-valueCardiovascular HR (95% CI)P-valueCancer HR (95% CI)P-value
  1. HR indicates hazard ratio per SD change in galectin-3. *Multivariable model included adjustment for age, gender, hypertension, hypercholesterolaemia, diabetes and smoking. **Multivariable full adjustment included all variables shown in Table 1.

Unadjusted1.46 (1.37–1.56)0.0001.56 (1.39–1.75)0.0001.41 (1.28–1.56)0.000
Age and gender adjusted1.12 (1.03–1.21)0.0071.16 (1.01–1.35)0.0441.10 (0.97–1.23)0.128
Multivariable classical risk factor adjusted*1.09 (1.01–1.19)0.0361.10 (0.94–1.28)0.2281.08 (0.96–1.22)0.222
Multivariable classical risk factor adjusted* + hsCRP1.09 (1.00–1.19)0.0441.08 (0.92–1.27)0.3281.09 (0.96–1.23)0.171
Multivariable classical risk factor adjusted* + UAE1.08 (1.00–1.18)0.0591.09 (0.93–1.27)0.2821.07 (0.95–1.21)0.242
Multivariable classical risk factor adjusted* + NT-proBNP1.03 (0.95–1.12)0.4530.93 (0.80–1.08)0.3631.08 (0.95–1.21)0.241
Multivariable full adjustment**1.02 (0.92–1.12)0.7700.91 (0.75–1.12)0.3811.08 (0.94–1.25)0.248

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

To our knowledge, this is the first study to evaluate plasma galectin-3 levels in the general population. We observed that plasma galectin-3 is strongly correlated with classical risk factors of CV disease, including smoking, diabetes, blood pressure, serum lipids, BMI, kidney function, hs-CRP and NT-proBNP. Furthermore, we observed a strong association between galectin-3 and age and gender. Galectin-3 values were higher in female compared with male subjects and were elevated with increasing age. We observed a remarkable gender interaction between galectin-3 and most risk factors for CV disease. Finally, increased levels of galectin-3 are associated with increased risk of all-cause mortality in the general population, independent of classical risk factors of CV disease.

Galectin-3 was discovered about 15 years ago and is thought to be involved in inflammation and fibrosis. Galectin-3 is widely distributed throughout the body, including being expressed in heart, brain and blood vessels [2]. In particular, secretion of galectin-3 is associated with activation of macrophages and fibroblasts, which leads to inflammation and fibrosis. Fibrosis is generally considered to be vital for tissue repair, including cardiac and renal disease [19–22]. It has been reported that galectin-3 is produced by macrophages and activated myofibroblasts in experimental models of cardiac [23, 24], renal [25, 26] and liver fibrosis [27].

A pathophysiological role of galectin-3 in the development and progression of HF has been described, and it is supported by observations in several experimental models of HF, such as myocarditis, cardiomyopathy, diabetes and angiotensin II-induced hypertension (reviewed in [9]). Furthermore, increased plasma galectin-3 levels have been reported in (metastasized) cancer [4, 5, 13, 14, 28–30], Behçet’s disease [7] and obesity [8], and galectin-3 levels may even guide therapy [14]. We hypothesize that deleterious signals that precede overt disease initiate subclinical tissue changes associated with galectin-3 activation and secretion. Further mechanistic studies should be conducted to unravel how galectin-3 is activated and secreted. This may have clinical importance, as galectin-3-targeted therapy has been proposed as a therapeutic option in diseases associated with increased galectin-3 levels [30].

To date, the potential clinical importance of plasma galectin-3 in CV disease has predominantly been studied in cohorts with HF [10–12]. Galectin-3 may accelerate adverse ventricular remodelling [9], and remodelling is a dominant determinant of HF outcome [31]. Only one report has been published describing galectin-3 distribution in the general population. Christenson et al. [17] measured galectin-3 levels in 1092 subjects above 55 years of age who were free of established CV disease. No clinical follow-up was recorded, and interactions with clinical and biochemical measures, such as BMI, blood pressure, kidney and liver function, and blood glucose, were not reported. The present analysis, in almost 8000 subjects, has superior statistical power and clearly shows a correlation between galectin-3 and risk factors of CV disease, as well as an important relationship with gender and age.

Here, we provide further insight into the clinical and biochemical correlates of galectin-3, and its potential usefulness as a biomarker. The PREVEND cohort is characterized by a wide age range (28–75 years) and has almost 80 000 subject-years of follow-up. Therefore, this is by far the largest cohort in which galectin-3 has been measured. We confirm that galectin-3 levels are higher in women than in men. Furthermore, we show that galectin-3 biology in the general population may be distinctly different in women and men. The correlation between galectin-3 and various factors, such as blood pressure, serum lipids and BMI, is substantially and significantly different in female and male subjects. It is possible that regulation of galectin-3 is modulated by sex hormones, but currently it remains unclear how these intriguing observations can be explained. Prognostic value, however, seems equally strong for male and female subjects, so that gender-specific elevations carry equal prognostic importance. The gender interaction warrants further study, as for instance ventricular remodelling differs considerably between male and female subjects [32], and the role of galectin-3 in this may be of importance.

Another striking observation is the strong relationship between galectin-3 and age. Galectin-3 gradually increases with age, similar to most CV biomarkers. For example, different normal values for (NT-pro) BNP have been identified for different age strata [33]. We have obtained preliminary data in octogenarians, showing that galectin-3 levels were oftentimes higher than 20 ng mL−1 in subjects in whom HF was excluded according to current guidelines [33]. Although the current study was not designed to investigate this issue, we postulate that individual galectin-3 levels should be considered with respect to the age of the subjects.

Here, we report that elevated galectin-3 levels are associated with increased risk of mortality. Unadjusted analyses showed that, in particular, subjects in the highest quintile of galectin-3 have an increased risk of mortality. As expected, adjustment for gender and age substantially reduced the hazard ratio. Further correction for classical risk factors for CV disease did not essentially affect the hazard ratio, but galectin-3 lost its prognostic value after correction for NT-proBNP. Future studies should address the role of galectin-3 as a biomarker, alone or in combination with other biomarkers such as NT-proBNP and hs-CRP.

Elevations (minor or otherwise) in galectin-3 levels may indicate early, subclinical disease, including CV disease and cancer, and this may have translated into impaired survival over the course of our trial. The current findings may explain, at least in part, the pathways by which galectin-3 is activated. As hypertension, hypercholesterolaemia and kidney dysfunction are all clearly associated with myocardial and renal fibrosis, we hypothesize that these risk factors may confer their profibrotic effects via galectin-3. Although galectin-3 levels correlated with most established risk factors of CV disease, including blood pressure, serum lipids, renal function and hs-CRP, correlation coefficients were generally weak, suggesting that these factors only have a marginal impact on galectin-3 levels. It is possible that there are additional, as yet unknown, factors that lead to galectin-3 production and release. Although some reports have been published that describe the regulation of galectin-3 [34–36], there is still a lack of knowledge of the transcriptional and translational regulation, the mechanism(s) responsible for transport from the nucleus into the cytoplasm and how galectin-3 affects the entire body, as well as its half-life and clearance.

Study strengths and limitations

Strengths of this study include the community-based cohort with a large sample size and long follow-up, the blinded measurements of galectin-3 levels and adjustment for multiple confounders in multivariable analyses. However, several limitations should be acknowledged. First, circulating biomarkers may not adequately reflect the activation of a biomarker in tissues. Second, only baseline and not serial galectin-3 levels were measured. Third, our sample exclusively comprised Caucasian subjects, and it is therefore unknown whether our findings can be generalized to individuals of other ethnicities. Fourth, the PREVEND cohort was enriched for microalbuminuria; this should be considered when applying the results to other cohorts. Finally, we used plasma samples that were frozen in 1997 and 1998 at −80 °C. Currently, there is a lack of knowledge regarding the potential degradation of galectin-3 during long-term storage, although it is reassuring to note that the reported values are comparable with those obtained from analyses performed in fresh plasma [17].

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

The results of this study show that plasma galectin-3 levels are correlated with established risk factors for CV disease and mortality in subjects from the general population. Furthermore, plasma galectin-3 levels are higher in women compared with men, and there is a strong gender-specific interaction in the correlations between galectin-3 and other risk factors for CV disease. Increased plasma galectin-3 levels independently predict all-cause mortality. Additional cross-sectional and longitudinal studies are warranted to confirm the link between galectin-3 and risk factors for CV disease and establish its role as a potential risk marker in the prediction of future (CV) events and mortality.

Conflict of interest statement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

BG Medicine, Inc. has certain rights related to galectin-3 measurements. BG Medicine, Inc. provided an unrestricted research grant to the Department of Cardiology of the University Medical Center Groningen. Drs. van Veldhuisen and de Boer have received consultancy and speaker’s fees from BG Medicine, Inc.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

BG Medicine, Inc. (Waltham, MA, USA) supported this study by providing kits and technical support to perform the ELISAs. Dr. de Boer is supported by the Netherlands Heart Foundation (grants 2007T046) and the Innovational Research Incentives Scheme of the Netherlands Organization for Scientific Research (NWO VENI, grant 916.10.117).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Conflict of interest statement
  9. Acknowledgements
  10. References
  11. Supporting Information

Figure S1. Strong age-related increase in plasma galectin-3 is observed.

Figure S2. Graphical depiction of the relationship between plasma galectin-3 and BMI (a), LDL cholesterol (b), systolic blood pressure (c), and serum triglycerides (d).

Table S1. Baseline characteristics: gender-stratified quintiles according to galectin-3 levels.

Table S2. Number of times each variable is selected after 1000 bootstrap samples.

Table S3. Gender-stratified corrected * estimates added to basic model.

FilenameFormatSizeDescription
JOIM_2476_sm_TableS1-3-FigS1-2.doc197KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.