The authors have declared that no competing interests exist.
Conceived and designed the experiments: JRL WHL RLP. Performed the experiments: TU JRL WHL GW KZ EML JB RLP. Analyzed the data: JRL WHL GW RLP. Contributed reagents/materials/analysis tools: TU JRL WHL GW KZ EML JB RLP. Wrote the paper: TU JRL WHL GW KZ EML JB RLP.
‡ These authors are co-first authors on this work.
Data on the predictive role of estimated glomerular filtration rate (eGFR) and osteoprotegerin (OPG) for cardiovascular (CVD) and all-cause mortality risk have been presented by our group and others. We now present data on the interactions between OPG with stage I to III chronic kidney disease (CKD) for all-cause and CVD mortality.
The setting was a 15-year study of 1,292 women over 70 years of age initially randomized to a 5-year controlled trial of 1.2 g of calcium daily. Serum OPG and creatinine levels with complete mortality records obtained from the Western Australian Data Linkage System were available. Interactions were detected between OPG levels and eGFR for both CVD and all-cause mortality (P < 0.05). Compared to participants with eGFR ≥60ml/min/1.73m2 and low OPG, participants with eGFR of <60ml/min/1.73m2 and elevated OPG had a 61% and 75% increased risk of all-cause and CVD mortality respectively (multivariate-adjusted HR, 1.61; 95% CI, 1.27-2.05; P < 0.001 and HR, 1.75; 95% CI, 1.22-2.55; P = 0.003). This relationship with mortality was independent of decline in renal function (P<0.05). Specific causes of death in individuals with elevated OPG and stage III CKD highlighted an excess of coronary heart disease, renal failure and chronic obstructive pulmonary disease deaths (P < 0.05).
The association between elevated OPG levels with CVD and all-cause mortality was more evident in elderly women with poorer renal function. Assessment of OPG in the context of renal function may be important in studies investigating its relationship with all-cause and CVD mortality.
An inverse association between estimated glomerular filtration rate (eGFR) and cardiovascular disease (CVD) and all-cause mortality has been established in several populations [
Osteoprotegerin (OPG) is a glycoprotein that has been shown to regulate several important organ systems by inhibiting the interaction between receptor activator of nuclear factor kappa-B ligand (RANKL) and its receptor RANK [
While circulating OPG levels are increased in CKD patients [
We therefore sought to investigate both the pathophysiology and the predictive value of elevated OPG levels in elderly women with mild to moderate renal dysfunction.
At baseline written informed consent was obtained from all participants for the study and follow up of electronic health records. The Human Ethics Committee of the University of Western Australia approved the study protocol and consent form (approval number 05/06/004/H50). The Human Research Ethics Committee of the Western Australian Department of Health (DOHWA HREC) also approved the data linkage study (approval number #2009/24).
The participants were recruited in 1998 to a 5-year prospective, randomised, controlled trial of oral calcium supplements to prevent osteoporotic fractures, the Calcium Intake Fracture Outcome study (CAIFOS) [
Baseline medical history including the presence of diabetes, hypertension, smoking history (current smokers/former smokers or non-smokers) and medication use were recorded. Participant’s previous medical history and current medications were verified by their General Practitioner where possible. These data were coded using the International Classification of Primary Care—Plus (ICPC-Plus) method [
Fasting blood samples for biochemistry were collected in 1998 and 2003 with sera stored at -80°C until analysis. Free OPG was measured in 2005 using the baseline sera from 1,333 (89%) participants using a validated enzyme immunoassay (R&D Systems, Minneapolis, MN, USA) as previously described [
Prevalent disease status were derived from the International Classification of Diseases, Injuries and Causes of Death Clinical Modification (ICD-9-CM) [
Mortality records were obtained from WADLS for each study participant between 1998 and 2013. International Classification of Diseases, Injuries and Causes of Death (ICD) primary and multiple cause of death were determined from the coded death certificate using information in Parts 1 and 2 of the death certificate or all diagnosis text fields from the death certificate where ICD 10 coded death data were not yet available. Deaths were defined using diagnosis codes from the ICD: Clinical Modification (ICD-9-CM) [
Baseline characteristics are presented as mean ± SD for continuous variables or median and interquartile range (IQR) for non-normally distributed variables. OPG was not normally distributed and was log transformed for analyses. OPG levels were categorised as above and below median cut-point of 2.2ng/mL. Effect modification between covariates and elevated OPG with vascular and all-cause mortality was examined by interaction tests with significant interactions detected using Cox regression. Participants were then categorised into 4 groups according to their OPG levels (above the median; 2.2 ng/mL—elevated, below the median—low) and eGFR measured by CKD-EPI eGFR (≥ 60 mL/min/1.73m2 and < 60ml/min/1.73m2). Models adjusting for 5-year change in eGFR excluded individuals with loss to follow-up due to withdrawal from the study and/or death or no measurement of 5-year creatinine (n = 325). Unadjusted and multivariable- adjusted Cox regression analyses were undertaken using IBM SPSS Statistics Version 21 (2012, Armonk, NY: IBM Corp). No violations of the Cox proportional hazards assumptions were detected. To exclude the possibility of reverse causality additional analyses were undertaken excluding participants who died within the first 24 months. Multivariable-adjustments included baseline age, body mass index, smoking history, treatment code (calcium or placebo), hormone replacement therapy and comorbidity score. P-values of less than 0.05 in two tailed testing were considered statistically significant.
As we sought to determine whether eGFR modified or accounted for the previously observed relationship between OPG and mortality, interaction tests between OPG and eGFR were undertaken. Using both variables as continuous (log transformed OPG and eGFR ml/min/1.73m2) there was an interaction observed between OPG and eGFR for all-cause mortality (P = 0.044) and CVD mortality (P = 0.028). Similarly when eGFR was dichotomized into above or below stage 3 CKD (eGFR < 60 ml/min/1.73m2) there was a significant interaction with log transformed OPG and stage of CKD for all-cause mortality (P = 0.043) and CVD mortality (P = 0.016) or when dichotomizing by median OPG levels with eGFR (ml/min/1.73m2) the interaction term remained significant for all-cause (P = 0.021) and CVD mortality (P = 0.047). Accordingly the participants were stratified into 4 groups according to median OPG levels (<2.2 ng/mL and ≥2.2 ng/mL) and eGFR dichotomised by the presence or absence of stage III CKD (< and ≥60mL/min/1.73m2). Graphical representation of these interactions are presented in
Baseline characteristics of the four groups are presented in
Low OPG eGFR ≥60 | Elevated OPG eGFR ≥60 | Low OPG eGFR <60 | Elevated OPG eGFR <60 | |
---|---|---|---|---|
1.8 [1.5–2.0] | 2.7 [2.4–3.1] | 1.8 [1.6–2.0] | 2.8 [—2.0] | |
74.1 ± 8.6 | 73.2 ± 8.7 | 52.1 ± 6.4 | 50.9 ± 7.1 | |
74.6 ± 2.5 | 75.3 ± 2.6 | 75.2 ± 2.7 | 76.1 ± 3.0 | |
27.2 ± 4.4 | 26.8 ± 4.6 | 27.5 ± 4.3 | 27.7 ± 5.5 | |
242 (52.7) | 224 (54.1) | 81 (42.9) | 123 (53.9) | |
170 (37.2) | 151 (36.7) | 78 (40.8) | 76 (33.8) | |
20 (4.4) | 34 (8.2) | 11 (5.8) | 22 (9.6) | |
2 (0.4) | 5 (1.2) | 6 (3.1) | 10 (4.4) | |
32 (7.0) | 32 (7.7) | 20 (10.5) | 34 (14.9) | |
5 (1.1) | 13 (3.1) | 10 (5.2) | 13 (5.7) | |
2 (0.5) | 1 (0.2) | 0 (0.0) | 1 (0.4) | |
74 (16.2) | 65 (15.8) | 34 (17.8) | 31 (13.7) | |
136.2 ± 17.0 | 139.7 ± 18.5 | 138.0 ± 20.5 | 138.5 ± 17.1 | |
72.4 ± 10.3 | 73.6 ± 10.6 | 73.9 ± 12.0 | 72.8 ± 11.6 | |
93.6 ± 10.9 | 95.7 ± 11.7 | 95.2 ± 13.5 | 94.7 ± 11.8 | |
226 ± 43 | 226 ± 41 | 226 ± 40 | 231 ± 46 | |
141 ± 38 | 143 ± 38 | 141 ± 37 | 145 ± 43 | |
58 ± 15 | 56 ± 14 | 54 ± 13 | 54 ± 14 | |
131 ± 64 | 134 ± 56 | 156 ± 77 | 155 ± 71 | |
164 (35.7) | 166 (40.1) | 93 (48.7) | 129 (56.6) | |
71 (15.5) | 80 (19.3) | 46 (24.1) | 49 (21.5) | |
171 (15.5) | 93 (22.5) | 42 (22.0) | 68 (29.5) | |
1 [0–1.0] | 1 [0–2.0] | 1 [0–2.0] | 1 [0–2.0] |
Data expressed as mean ± SD or number and (%). Abbreviations: OPG, osteoprotegerin; CKD, chronic kidney disease; mmHg, millimeters mercury; CKD-EPI eGFR, Chronic Kidney Disease Epidemiology Collaboration estimated glomerular filtration rate.
* Significantly different between groups by ANOVA or χ2 test where appropriate (P<0.05)
† measured in 1,252 participants and
‡ measured in 974 participants.
In participants with an eGFR <60mL/min/1.73m2 (n = 419) per SD increase in OPG there was a 28% increase in the risk of 15 year death in unadjusted (HR 1.28, 95%CI 1.13–1.46, P<0.001) that remained significant after multivariable-adjustment (HR 1.25, 95%CI 1.09–1.43, P = 0.001). In participants with an eGFR ≥60mL/min/1.73m2 per SD increase in OPG there was a marginally non-significant 11% increase in the risk of 15 year death in unadjusted (HR 1.11, 95%CI 1.00–1.22, P = 0.055) that remained non-significant after multivariable-adjustment (HR 1.05, 95%CI 0.94–1.17, P = 0.398).
Participants with below-median OPG levels and eGFR ≥60mL/min/1.73m2 had the lowest incidence of all-cause mortality (35.3%) and therefore were designated the referent group.
Participants with above-median OPG levels and eGFR ≥60mL/min/1.73m2 had significantly higher incidence of 15-year all-cause mortality (44.0%) in unadjusted models (HR 1.35, 95%CI 1.09–1.66, P = 0.006) and that became non-significant after multivariable adjustment (
Multivariable-adjustments were baseline age, body mass index, smoking history, history of hormone replacement therapy, treatment code (calcium or placebo) and comorbidity score.
In participants with an eGFR <60mL/min/1.73m2 (n = 419) per SD increase in OPG there was a 43% increase in the risk of 15 year cardiovascular death in unadjusted (HR 1.43, 95%CI 1.19–1.71, P<0.001) that remained significant after multivariable-adjustment (HR 1.39, 95%CI 1.14–1.70, P = 0.001). In participants with an eGFR ≥60mL/min/1.73m2 per SD increase in OPG there was no increase in the risk of 15 year cardiovascular death in unadjusted (HR 1.06, 95%CI 0.89–1.26, P = 0.523) that remained non-significant after multivariable-adjustment (HR 1.00, 95%CI 0.83–1.12, P = 0.974).
Compared to participants with below-median OPG levels and eGFR ≥60mL/min/1.73m2, only participants with above-median OPG levels and eGFR <60mL/min/1.73m2 had significantly higher risk of 15-year CVD mortality before (HR 2.33, 95%CI 1.62–3.33, P < 0.001) and after adjustment for age, BMI, treatment, smoking history and prevalent renal disease and diabetes (
Multivariable-adjustments were baseline age, body mass index, smoking history, history of hormone replacement therapy, treatment code (calcium or placebo) and comorbidity score.
Unlike the primary cause of CVD death data participants with above-median OPG levels and eGFR ≥60mL/min/1.73m2 and an eGFR of <60mL/min/1.73m2 had increased risk of CVD-related deaths compared to participants with below-median OPG levels and eGFR ≥60mL/min/1.73m2 before (HR 1.46, 95%CI 1.11–1.91, P = 0.006 and HR 2.45, 95%CI 1.84–3.25, P < 0.001 respectively) and after multivariable adjustment (
Low OPG eGFR ≥60 | Elevated OPG eGFR ≥60 | Low OPG eGFR <60 | Elevated OPG eGFR <60 | P value | |
---|---|---|---|---|---|
51 (11.1) | 54 (13.1) | 16 (8.4) | 26 (11.5) | 0.405 | |
41 (8.9) | 26 (6.3) | 19 (9.9) | 26 (11.5) | 0.130 | |
6 (1.3) | 5 (1.2) | 2 (1.0) | 6 (2.7) | 0.441 | |
22 (4.8) | 33 (8.0) | 9 (4.7) | 19 (8.4) | 0.106 | |
34 (7.4) | 43 (10.4) | 17 (8.8) | 27 (11.9) | 0.220 | |
65 (14.2) | 58 (14.0) | 18 (9.4) | 38 (16.8) | 0.184 | |
31 (6.8) | 33 (8.0) | 14 (7.3) | 28 (12.4) | 0.081 | |
15 (3.3) | 21 (5.1) | 4 (2.1) | 12 (5.3) | 0.197 | |
27 (5.9) | 27 (6.5) | 8 (4.2) | 16 (7.1) | 0.619 | |
10 (2.2) | 14 (3.4) | 4 (2.1) | 5 (2.2) | 0.641 |
Data expressed as mean ± SD or number and (%). P value represents overall P value for the trend by χ2 test. Abbreviations: OPG, osteoprotegerin; CKD, chronic kidney disease; CVD cardiovascular disease; CHD coronary heart disease, COPD, chronic obstructive pulmonary disease.
*Disease specific multiple cause of deaths numbers do not add up to total mortality as multiple causes of death are possible.
No. (%) | HR (95%CI) |
P value | |
---|---|---|---|
96/459 (20.9) | 1.00 (reference) | ||
116/414 (28.0) | 1.27 (0.97–1.67) | 0.086 | |
49/191 (25.7) | 1.06 (0.75–1.50) | 0.743 | |
47/459 (10.2) | 1.00 (reference) | ||
50/414 (12.1) | 1.09 (0.73–1.63) | 0.690 | |
25/191 (13.1) | 1.04 (0.64–1.71) | 0.864 | |
11/459 (2.4) | 1.00 (reference) | ||
10/191 (5.2) | 1.81 (0.76–4.31) | 0.179 | |
25/459 (5.4) | 1.00 (reference) | ||
33/414 (8.0) | 1.43 (0.84–2.42) | 0.185 | |
9/191 (4.7) | 0.79 (0.37–1.70) | 0.540 | |
Abbreviations: OPG, osteoprotegerin; HR, hazard ratio; Total CVD, total cardiovascular disease; CHD, coronary heart disease; COPD, chronic obstructive pulmonary disease.
*Multivariable-adjustments were baseline age, body mass index, smoking history, history of hormone replacement therapy, treatment code (calcium or placebo) and comorbidity score.
To assess whether the relationship between circulating OPG levels with mortality outcomes was attributed to long-term renal decline, a Cox regression model including the 5-year change in eGFR was created (
Multivariable-adjustments included 5-year change in CKD-EPI eGFR (n = 970), age, body mass index, smoking history, history of hormone replacement therapy, treatment code (calcium or placebo) and comorbidity score.
When excluding participants with CVD at baseline (n = 302) only participants with elevated OPG and an eGFR <60 ml/min/1.73m2 were at an increased risk of CVD death before and after multivariate-adjustment (HR 2.09–1.33–3.29, P = 0.001 and HR 1.59, 95%CI 1.00–2.54, P = 0.050 respectively). Participants with elevated OPG and an eGFR ≥60 ml/min/1.73m2 had an increased risk of all-cause mortality before multivariable-adjustment (HR 1.30–1.03–1.66, P = 0.037) but not after multivariable-adjustment (HR 1.20–0.94–1.54, P = 0.153) while those with elevated OPG and an eGFR <60 ml/min/1.73m2 were at an increased risk before and after multivariate-adjustment (HR 1.99–1.50–2.62, P < 0.001 and HR 1.65, 95%CI 1.23–2.19, P = 0.001 respectively). To assess potential reverse causality bias we excluded all deaths that occurred within the first 24 months of study with little change to the overall result (elevated OPG and ≥60 ml/min/1.73m2 HR 1.22, 95%CI 0.98–1.52, P = 0.069 and elevated OPG and <60 ml/min/1.73m2 HR 1.63, 95%CI 1.28–2.08, P < 0.001 respectively) and CVD mortality (elevated OPG and <60 ml/min/1.73m2 HR 1.75, 95%CI 1.20–2.55, P = 0.004).
In a cohort of community-based ambulant elderly women, the association between OPG levels and all-cause and cardiovascular mortality appeared to be modified by reduced kidney function. Women with above-median OPG levels and an eGFR <60mL/min/1.73m2 had a 71–88% increased risk of all-cause and cardiovascular mortality and a 4-fold increase in renal failure mortality compared to participants with below-median OPG levels and without moderate CKD. In addition to demonstrating an association between OPG and mortality in individuals with stage III CKD, this study raises important considerations of how OPG may be involved in the pathophysiological of CVD mortality. These findings extend upon the previous study by our group [
The increased mortality risk identified in individuals with elevated OPG levels appeared to diverge between 3 to 5 years after circulating OPG assessment and continued to diverge until the end of follow-up, suggesting a single measurement of circulating OPG levels may potentially help identify individuals with progressive disease that could result in death over a 15-year period [
Regarding potential mechanisms, OPG is derived locally from both bone and vascular smooth muscle cells and is present in high concentrations throughout all layers of normal and atherosclerotic blood vessel walls [
Although it has been suggested that the elevated OPG levels detected in CKD and ESRD patients are attributable to reduced renal clearance, it is also possible that an excess of production of OPG could also be contributory given its role in regulating calcification. However the rapid fall observed in OPG levels following successful renal transplantation in ESRD patients supports the former hypothesis [
Using Multiple Cause of Death (MCoD) data, we were able to identify coronary artery disease as the major identifiable cause of increased cardiovascular disease mortality in participants with an eGFR <60/ml/min/1.73m2. The clinical manifestations of coronary artery disease include: myocardial infarction, angina and acute coronary syndrome and chronic coronary heart disease. Given the relationship between OPG and long-term renal decline observed in this cohort previously [
There are several limitation of this study that must be considered, firstly these findings are only in elderly women and as such may not be generalizable to men and other younger age groups. A further limitation of the current study is that we did not test the interaction between OPG and other measures of chronic kidney disease such as albuminuria and did not adjust for other markers of inflammation, atherosclerosis or vascular calcification. Also as this was a prospective cohort study we lack data on the temporal sequence of the changes in OPG levels and the development of clinical events and thus cannot identify causality. However despite the lack of a temporal sequence of the change in circulating OPG we have demonstrated that OPG measured at a single time point when applied in the context moderate CKD identifies patients with a poorer clinical prognosis for up to 15 years in this cohort suggesting measuring OPG in individuals with CKD may improve identification of individuals with a poorer prognosis.
The strengths of this study include the complete and accurate data collection over a 15-year period independent of self-report in a large cohort of subjects and the ability to accurately examine the association between baseline OPG levels and renal function with long-term clinical outcomes removing the possibility of recall or retention bias. This comprehensive outcome data in conjunction with the measurement of both circulating OPG and creatinine has allowed the identification of the interaction between renal dysfunction and elevated OPG with CVD and all-cause mortality. In addition the use of multiple cause of death data which highlights relationships between concurrent disease processes and allows patterns of pathological damage to be elucidated is a strength of the study and has identified both renal failure and chronic obstructive pulmonary disease as potential mechanisms underlying the association between OPG and all-cause mortality.
In conclusion our findings supports the concept that the increased CVD and all-cause mortality risk associated with elevated OPG levels is primarily seen in individuals with moderate CKD. Measuring circulating OPG levels in individuals with at least moderate CKD may allow early identification of those at higher risk of CVD and all-cause mortality.
The authors wish to thank the staff at the Data Linkage Branch, Hospital Morbidity Data Collection and Registry of Births, Deaths and Marriages for their work on providing the data for this study.