- Email: [email protected]
Association of chronic kidney disease categories defined with different formulae with major adverse events in patients with peripheral vascular disease
Atherosclerosis 232 (2014) 289e297
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Association of chronic kidney disease categories defined with different formulae with major adverse events in patients with peripheral vascular disease Jonathan Golledge a, b, *, Carla Ewels a, c, Reinhold Muller a, c, Phillip J. Walker d, e a
Queensland Research Centre for Peripheral Vascular Disease, School of Medicine and Dentistry, James Cook University, Townsville, Australia Department of Vascular and Endovascular Surgery, The Townsville Hospital, Townsville, Australia School of Public Health, Tropical Medicine and Rehabilitation Sciences, James Cook University, Townsville, Australia d University of Queensland School of Medicine, Discipline of Surgery and Centre for Clinical Research, Brisbane, Australia e Department of Vascular Surgery, The Royal Brisbane and Women’s Hospital, Brisbane, Australia b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 September 2013 Received in revised form 1 November 2013 Accepted 1 November 2013 Available online 2 December 2013
Objective: The aim of this study was to compare the ability of eGFR calculated by modification of diet in renal disease (MDRD), Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) and Lund-Malmö formulae in predicting major adverse events in peripheral vascular disease (PVD) patients. Methods: We prospectively recruited 2137 patients, measured serum creatinine to calculate eGFR using three different formulae and grouped patients into eGFR categories 90, 60e89, 45e59, 30e44, 15e29 and <15 ml/min/1.73 m2. Patients were followed up for a median of 1.3 (inter-quartile range 0.3e3.6) years. The primary outcome was the combined incidence of myocardial infarction, stroke or death. The ability of eGFR categories defined with the different formulae to predict outcome was assessed using the net reclassification index. Results: 1450 (67.9%), 1515 (70.9%) and 1813 (84.8%) patients had eGFR <90 ml/min/1.73 m2 according to the CKD-EPI, MDRD and Lund-Malmö formulae, respectively. Using the CKD-EPI formula 276 (12.9%) patients were reclassified to a different eGFR category in comparison to the MDRD formula and the prediction of outcome was improved (net reclassification index 0.106, p < 0.001). Using the Lund-Malmö formula 563 (26.3%) patients were reclassified to a different eGFR category in comparison to the MDRD formula and the prediction of outcome was improved (net reclassification index 0.108, p < 0.001). Classification using the CKD-EPI and Lund-Malmö formulae was equally effective at predicting outcome (net reclassification index - 0.002, p ¼ 0.397). Conclusions: eGFR categories determined with the CKD-EPI and Lund-Malmö formulae are equally effective at predicting major adverse events in patients with PVD. Ó 2013 Elsevier Ireland Ltd. All rights reserved.
Keywords: Peripheral vascular disease Chronic kidney disease Estimated glomerular filtration rate
1. Introduction The prevalence of chronic kidney disease (CKD) is estimated to be >10% and increasing [1]. The prevalence of CKD is particularly high in some patients groups, such as patients with cardiovascular disease [2]. Peripheral vascular diseases (PVD) are a group of conditions affecting the vessels outside the heart [3e6]. CKD is a risk factor for PVD development and CKD patients with PVD have an increased incidence of major cardiovascular events [7,8]. There is
* Correspondence author. Queensland Research Centre for Peripheral Vascular Disease, School of Medicine and Dentistry, James Cook University, Townsville 4811, Australia. Tel.: þ61 7 4796 1417; fax þ61 7 4796 1401. E-mail address: [email protected] (J. Golledge). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.11.034
current controversy over how CKD is best defined with a large number of different ways to estimate kidney function by calculating estimated glomerular filtrate rate (eGFR) being available [9,10]. A number of previous studies have associated CKD with major adverse events in patients with PVD [11e20]. In these studies a number of different equations have been used to calculate eGFR, including the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) and the modification of diet in renal disease (MDRD) formulae [11e20]. Recently eGFR calculated using the Lund-Malmö formula has been suggested as being more accurate in some patients groups [21]. There are a number of possible reasons to suspect that the most appropriate formula to assess eGFR in PVD patients might be different from healthy individuals. The prevalence of CKD is high amongst PVD patients and eGFR formulae vary in their ability to estimate severe renal function impairment
290
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
[1,2,9,10]. Furthermore, the aetiology of CKD in PVD is more likely to be related to renovascular disease which may also influence the most appropriate formula to use [1,2,9,10]. Currently it is not clear which formula best predicts major adverse events in patients with PVD. The primary aim of this study was to assess which of three formulae used to calculate eGFR best predicted major adverse events in patients with PVD. 2. Methods 2.1. Study design This study was designed as an on-going prospective cohort investigation of patients with PVD aimed at assessing risk predictors of PVD presence and outcome commencing in 2002, as previously described [22]. Since the comparative ability of different eGFR formulae to predict major adverse events in PVD patients had not been previously examined sample size calculations were not straight forward and no formal calculation was performed [11e20]. Monte-Carlo simulations suggest that a multivariate regression model is powered sufficiently when 10 outcome events per degree of freedom of the predictor variables are observed [23]. We estimated that the combined incidence of myocardial infarction, stroke or death at one year would be approximately 10% and planned to adjust for 11 variables (age, male sex, hypertension, diabetes, smoking, coronary heart disease (CHD), presenting complaint, statin prescription, aspirin prescription, angiotensin converting enzyme (ACE) inhibitor prescription and angiotensin receptor blocker (ARB) prescription) in our regression model. Based on these estimates we felt that a sample size of over 2000 patients would be well powered to examine the association of eGFR categories with major adverse events. 2.2. Patients
This included telangiectasia or reticular veins (C1); varicose veins of 3 mm (C2); oedema (C3); skin changes due to chronic venous disease (C4); healed venous ulcer (C5); and active venous ulcer (C6). All patients underwent venous duplex imaging; b) Miscellaneous PVD problems including aortic dissection, reno-vascular hypertension, mesenteric ischaemia and peripheral vascular trauma were diagnosed based on history, examination and imaging using duplex imaging or computed tomographic angiography; c) Asymptomatic carotid artery stenosis: Defined as the presence of 50% stenosis or occlusion of at least one carotid artery identified by carotid duplex but the absence of physician confirmed symptoms of focal transient ischaemic attack, amaurosis fugax or stroke as previously described [24]; d) Mild lower limb or upper limb peripheral athero-thrombosis: This included patients with intermittent claudication, atypical or no symptoms with clinical evidence of lower or upper limb ischaemia but not critical lower limb ischaemia. Limb peripheral athero-thrombosis was confirmed by a vascular specialist by identification of absence of lower or upper limb pulses, ankle brachial pressure index <0.9 and/or significant stenosis (>50%) or occlusion of lower or upper limb arteries on computed tomographic angiography or duplex imaging [25,26]. e) Aneurysm of the aorta or peripheral arteries: Aortic aneurysm was defined as maximum aortic diameter 30 mm [25e27]. Iliac artery aneurysm was defined by common or internal iliac artery diameters 15 and 8 mm, respectively. Femoral artery aneurysm was defined by common femoral or superficial femoral artery diameter of 15 mm. Popliteal artery aneurysm was defined as popliteal artery diameter 9 mm as previously described [28]; f) Symptomatic carotid artery stenosis: Defined as the presence of 50% stenosis or occlusion of at least one carotid artery identified with carotid duplex with the presence of physician confirmed symptoms of focal transient ischaemic attack, amaurosis fugax or stroke as previously described [24]; g) Critical lower limb ischaemia: Rest pain, arterial ulcer or gangrene of the leg due to athero-thrombosis of the lower limb. Peripheral athero-thrombosis was confirmed as detailed above [25,26]. For patients with more than one presenting complaint classification was determined by the complaint which was deemed most severe.
Patients were recruited from in and out-patient vascular services at The Townsville Hospital, The Mater Hospital Townsville and The Royal Brisbane and Women’s Hospital. Patients with all types of PVD were considered for inclusion. All patients diagnosed as having any type of PVD by a Royal Australasian College of Surgeons accredited vascular specialist were considered for inclusion into the study. Inclusion criteria for the current study included a diagnosis of PVD, the assessment of serum creatinine to enable the calculation of eGFR and a least one follow-up assessment as an in or outpatient. Ethical approval for the study was granted by the local Institutional Ethics Committees at The Townsville Hospital, The Mater Hospital Townsville, The Royal Brisbane and Women’s Hospital and James Cook University. Written informed consent was obtained from participants.
Hypertension was defined by a history of high blood pressure or receiving treatment to reduce blood pressure [22,24e28]. Diabetes was defined by a fasting blood glucose concentration 7.0 mM, or history of, or treatment for hyperglycaemia [22,24e28]. Smoking status was classified as ever and never smokers [22,24e28]. CHD was defined by a history of myocardial infarction, angina or treatment for coronary artery disease [22,24e28].
2.3. Definition of presenting complaint
2.6. Medications
Presenting category was broadly defined into one of seven groups namely venous disease; miscellaneous PVDs (including aortic dissection, reno-vascular hypertension, mesenteric ischaemia and peripheral vascular trauma); asymptomatic carotid stenosis; mild lower limb or upper limb peripheral atherothrombosis; aneurysm of the aorta or peripheral arteries; symptomatic carotid artery stenosis; and critical lower limb ischaemia, as previously described in detail [22,24e28].
At the time of recruitment a list of each patient’s medications was recorded including whether the participants were prescribed statins, aspirin, ACE inhibitors or ARBs.
2.4. Definitions and diagnosis of PVD PVD was defined using the following criteria: a) Venous disease: This was defined according to the CEAP classification [29].
2.5. Definition of other risk factors
2.7. Measurement of serum creatinine and calculation of eGFR Serum creatinine was measured in a pathology laboratory using a spectrophotometry method in line with established guidelines as previously described [30,31]. eGFR was calculated using the CKD-EPI, MDRD (isotope dilution mass spectrometry aligned) and Lund-Malmö formulae [9,21,32]. All these formulae have been previously described in detail and utilise creatinine, age and gender in their calculations [9,21,32]. The following
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
291
eGFR categories were defined: 90, 60e89, 45e59, 30e44, 15e 29 and <15 ml/min/1.73 m2 as currently used to define CKD stages [1].
3. Results
2.8. Follow-up
The presenting complaints of the 2137 patients included aortic or peripheral aneurysms (n ¼ 712; 33.3%), venous disease (n ¼ 420; 19.7%), non-critical limb athero-thrombosis (n ¼ 365; 17.1%), symptomatic carotid artery stenosis (n ¼ 245; 11.5%), critical lower limb ischaemia (n ¼ 219; 10.2%), asymptomatic carotid artery stenosis (n ¼ 148; 6.9%), and miscellaneous PVDs (n ¼ 28; 1.3%). Sixty seven per cent of patients were male, 73% had a history of ever smoking, 67% had a history of hypertension, 40% a history of CHD and 24% had diabetes (Supplementary Table 1). Approximately half the patients were receiving a statin and aspirin, 35% receiving an angiotensin converting enzyme inhibitor and 20% receiving an angiotensin receptor blocker.
Patients were followed up through attendance at out-patient clinics and/or as an in-patient as part of their normal medical care. Patients with venous disease and miscellaneous PVD were followed up as needed depending on the severity of their presenting complaint and any medical intervention planned. Patients with limb athero-thrombosis and carotid artery stenosis were generally reviewed 6 months after their initial assessment and then yearly unless symptoms or imaging findings changed [23e25]. Patients with small aneurysms or large aneurysms that had been repaired were followed up yearly or 6 monthly if the aneurysm was nearing a diameter at which intervention was indicated [26,27].
3.1. Characteristics of the cohort
3.2. Relationship between risk factors and eGFR categories
The primary outcome was the combined incidence of myocardial infarction, stroke or death. Charts and hospital electronic records of all patients were reviewed at least once for the identification and date of any outcome events by a vascular specialist or research nurse. Where uncertainty about outcome data was present a discussion occurred between the research nurse and vascular specialist and a consensus was reached. Myocardial infarctions and strokes were defined based on diagnosis by a Royal Australasian College of Physicians accredited physician in line with international guidelines [33,34]. For patients that had not experienced the primary outcome follow-up was concluded, i.e. censoring occurred, at the date of last in or outpatient review.
eGFR calculated by the CKD-EPI formula (median 80.3, interquartile range, IQR, 60.8e93.3 ml/min/1.73 m2) was significantly higher than that estimated by both the MDRD equation (median 76.6, IQR 60.6e93.6 ml/min/1.73 m2 p ¼ 0.014, Wilcoxon Signed Rank Test) and Lund-Malmö (median 70.3, IQR 55.9e83.0 ml/min/ 1.73 m2 p < 0.001, Wilcoxon Signed Rank Test) formulae. Nine patients had end-stage renal failure requiring dialysis and had eGFR <15 ml/min/1.73 m2. Overall patients with eGFR 90 ml/min/ 1.73 m2 defined by the CKD-EPI formula were more likely to be younger, female, never smokers and present with venous disease (Supplementary Table 1). Patients with eGFR 90 ml/min/1.73 m2 were less likely to have hypertension, diabetes, CHD, or be prescribed statins, aspirin, ACE inhibitors and angiotensin receptor blockers (Supplementary Table 1). Similar associations were noted when eGFR was defined using the MDRD and Lund-Malmö formulae (data not shown).
2.10. Statistical analyses
3.3. Relationship between outcome and eGFR
Quantitative data were not normally distributed and therefore are presented as median and inter-quartile range and assessed by non-parametric tests. Nominal data are presented as number and percentages and compared by chi-squared test. The associations of eGFR and eGFR categories with the combined incidence of myocardial infarction, stroke or death were assessed using Kaplan Meier estimates and Cox proportional hazard analyses. Cox proportional hazard analysis was adjusted for age >68 years (median age of the cohort), male sex, hypertension, diabetes, smoking, CHD, presenting complaint, statin prescription, aspirin prescription, ACE inhibitor prescription and ARB prescription. For the Cox analyses presenting PVD problem was defined as an indicator from one to seven in the following order: venous disease; miscellaneous PVDs; asymptomatic carotid stenosis; mild lower limb or upper limb peripheral athero-thrombosis; aneurysm of the aorta or peripheral arteries; symptomatic carotid artery stenosis; and critical lower limb ischaemia. Other adjusted variables were defined as present or absent. The variables adjusted for were chosen as they are recognised determinants of outcome for patients with cardiovascular disease. The combined incidence of myocardial infarction, stroke or death was compared between groups of subjects in different eGFR categories using the log rank test. The ability of eGFR categories defined with the different formulae to predict the combined incidence of myocardial infarction, stroke or death was assessed using the net reclassification index which was calculated as previously described [35].
Patients were followed up for a median of 1.3 (inter-quartile range 0.3e3.6) years. During this period 106 patients had a myocardial infarction, 72 patients had a stroke and 327 patients died. 42 of the patients who suffered a myocardial infarction and 33 of those who experienced a stroke subsequently died during later follow-up. The primary outcome of myocardial infarction, stroke or death occurred in 425 independent patients while the other 1712 patients were censored at last follow-up. Overall the estimated combined incidences of myocardial infarction, stroke or death were 8.6, 14.2 and 21.5% at 1, 2 and 3 years, respectively. The estimated combined incidence of myocardial infarction, stroke or death at three years were 12.1, 19.6, 29.7, 34.7, 40.7 and 61.2% in patients that had an CKD-EPI eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1a). The estimated combined incidence of myocardial infarction, stroke or death at three years were 13.3, 19.1, 29.2, 34.0, 42.8 and 64.1% in patients that had an MDRD eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1b). The estimated combined incidence of myocardial infarction, stroke or death at three years were 9.1, 16.7, 27.2, 35.8, 39.7 and 61.2% in patients that had an Lund-Malmö eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1c). After adjustment for age, male gender, hypertension, diabetes, ever smoking, CHD, presenting problem and prescription of statins, aspirin, ACE inhibitors and ARBs a 10 ml/min/1.73 m2 increase in eGFR (approximately one standard deviation in the whole population) was
2.9. Recording of outcome data
292
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
Fig. 1. Kaplan Meier curves showing the incidence of myocardial infarction, stroke or death in relation to eGFR category at entry defined with the a) CKD-EPI; b) MDRD; and c) LundMalmö formulae. Vertical lines represent subjects censored at loss to follow-up.
associated with a reduced incidence of myocardial infarction, stroke or death although this finding was stronger for values estimated with the CKD-EPI and Lund-Malmö formulae (CKD-EPI RR 0.90, 95% CI 0.86e0.94, p < 0.001; MDRD RR 0.96, 95% CI 0.92e1.00, p ¼ 0.027; Lund-Malmö RR 0.88, 95% CI 0.84e0.93, p < 0.001). eGFR categories 30e44, 15e29 and <15 ml/min/1.73 m2 were independently associated with increased incidence of myocardial infarction, stroke or death when calculated using the CKD-EPI and MDRD formulae (Table 1). Only eGFR category <15 ml/min/1.73 m2 was independently associated with increased incidence of myocardial infarction, stroke or death when calculated using the Lund-Malmö formula (Table 1). 3.4. eGRF categories according to the different formulae and relationship with risk factors and outcomes 1450 (67.9%), 1515 (70.9%) and 1813 (84.8%) patients had eGFR <90 ml/min/1.73 m2 according to the CKD-EPI, MDRD and LundMalmö formulae, respectively. Using the CKD-EPI formula 113 (5.3%) patients were reclassified to a worse eGFR category (76 to eGFR 60e89; 18 to eGFR 45e59; 11 to eGFR 30e44; 6 to eGFR 15e 29 and 2 to eGFR <15 ml/min/1.73 m2) and 163 (7.6%) to a less advanced eGFR category (141 to eGFR 90; 20 to eGFR 60e89; 1 to eGFR 45e59; and 1 to eGFR 30e44 ml/min/1.73 m2) in comparison to classification using the MDRD formula. Thus most of the
reclassification was for patients who were categorised to have an eGFR 90 or 60e89 ml/min/1.73 m2 according to the MDRF formula. 76 of 622 (12.2%) patients classified to have an eGFR 90 with the MDRF formula were reclassified to have an eGFR of 60e89 ml/ min/1.73 m2 using the CKD-EPI formula. 141 of 994 (14.2%) patients classified to have an eGFR of 60e89 with the MDRF formula were reclassified to have an eGFR 90 ml/min/1.73 m2 using the CKD-EPI formula. The risk factors and outcomes of patients who had an eGFR 90 according to one or both formulae in relation to whether they were reclassified by the CKD-EPI formula are shown in Table 2. Patients reclassified to a worse eGFR category were older and more likely to be male, have hypertension, have diabetes, have CHD and present with an aneurysm or critical limb ischaemia. They were also more likely to be prescribed a statin, aspirin and an ACE inhibitor. Patients reclassified to a worse category had increased incidence of myocardial infarction, stroke or death while patients reclassified to a better eGFR category had reduced incidence of these events (Table 2, Supplementary Fig. 1). The net reclassification index improvement using the CKD-EPI as opposed to the MDRD formula was 0.106, p < 0.001. Using the CKD-EPI formula no patients were reclassified to a worse eGFR category but 589 (27.6%) were reclassified to a less advanced eGFR category (363 to eGFR 90; 146 to eGFR 60e89; 47 to eGFR 45e59; and 33 to eGFR 30e44 ml/min/1.73 m2) in comparison to classification using the Lund-Malmö formula. Generally
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297 Table 1 Association of eGFR categories calculated with CKD-EPI, MDRD and Lund-Malmö formulae with major adverse events (myocardial infarction, stroke or death) in 2137 patients with peripheral vascular disease eGFR category (ml/min/1.73 m2 )
Relative risk
Calculated with CKD-EPI formula 90 1.00 60e89 0.98 45e59 1.23 30e44 1.59 15e29 1.90 <15 3.38 Calculated with MDRD formula 90 1.00 60e89 0.97 45e59 1.21 30e44 1.48 15e29 2.06 <15 3.17 Calculated with Lund-Malmö formula 90 1.00 60e89 0.88 45e59 1.14 30e44 1.42 15e29 1.66 <15 3.14
95% CI
P value
Reference group 0.72e1.33 0.86e1.75 1.07e2.36 1.14e3.16 1.68e6.80
0.903 0.250 0.022 0.014 0.001
Reference group 0.74e1.27 0.88e1.66 1.02e2.15 1.28e3.29 1.49e6.78
0.843 0.249 0.037 0.003 <0.001
Reference group 0.56e1.39 0.69e1.86 0.84e2.43 0.95e2.93 1.44e6.84
0.575 0.614 0.194 0.078 0.004
eGFR ¼ Estimated glomerular filtration; CKD-EPI ¼ Chronic Kidney DiseaseEpidemiology Collaboration group formula; MDRD ¼ Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned formula; All models were adjusted for age above 68 years, male gender, coronary heart disease, hypertension, diabetes, ever smoker, presenting problem, statin prescription, aspirin prescription, angiotensin converting enzyme inhibitor prescription and angiotensin receptor blocker prescription.
patients reclassified to less severe eGFR categories were less likely to have established cardiovascular risk factors than those classified in the same eGFR category according to both formulae (Table 3). The relationship between reclassification using the CKD-EPI and the Lund-Malmö formulae and incidence of myocardial infarction, stroke or death are shown in Table 3 and Supplementary Fig. 2.
293
Patients reclassified from the 60e89 to 90 ml/min/1.73 m2 category by the CKD-EPI formula had a better prognosis than those in which eGFR was estimated to be 60e89 ml/min/1.73 m2 using both formulae. For other reclassifications the relationship was less clear (Table 3). The net reclassification index using the CKD-EPI as opposed to the Lund-Malmö formula was 0.002, p ¼ 0.397. Using the Lund-Malmö formula 551 (25.8%) patients were reclassified to a worse eGFR category (303 to eGFR 60e89; 150 to eGFR 45e59; 58 to eGFR 30e44; 38 to eGFR 15e29 and 2 to eGFR <15 ml/min/1.73 m2) and 12 (0.6%) to a less advanced eGFR category (5 to eGFR 90; 6 to eGFR 60e89; 1 to eGFR 45e59 ml/min/ 1.73 m2) in comparison to classification using the MDRD formula. Generally patients reclassified to more severe eGFR categories were more likely to have established cardiovascular risk factors than those classified as the same eGFR category according to both formulae (Table 4). The relationship between reclassification using the Lund-Malmö and the MDRD formulae and incidence of myocardial infarction, stroke or death is shown in Table 4. Patients reclassified from the 90 and 60e89 ml/min/1.73 m2 to more severe eGFR categories by the Lund-Malmö formula had a worse prognosis than those in which eGFR was the same using both formulae. For other reclassifications the relationship was less clear although the numbers of patients were small (Table 4). The net reclassification index improvement using the Lund-Malmö formula as opposed to the MDRD formula was 0.108, p < 0.001. 4. Discussion The current study suggests that the CKD-EPI and Lund-Malmö formulae are equally effective at defining patients with PVD into eGFR categories that are predictive of major adverse events. In contrast eGFR categories defined with the MDRD formula are inferior at predicting major adverse events for PVD patients. Our findings are in keeping with reports which have compared the use of the CKD-EPI and MDRD formulae in other populations, such as healthy older adults and patients presenting with heart
Table 2 Characteristics at recruitment and outcome of patients reclassified using the CKD-EPI formula compared to the MDRD formula. eGFR range MDRD (ml/min/1.73 m2) eGFR range CKD-EPI (ml/min/1.73 m2)
60e89 90
90 90
90 60e89
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem Venous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
141 56.3 (47.8e61.7) 68 (48.2%) 54 (38.3%) 19 (13.5%) 85 (60.3%) 26 (18.4%)
546 58.6 (48.7e66.0) 342 (62.6%) 258 (47.3%) 105 (19.2%) 358 (65.6%) 132 (24.2%)
76 76.2 (73.1e78.9) 59 (77.6%) 63 (82.9%) 23 (30.3%) 54 (71.1%) 38 (50.0%)
66 (46.8%) 3 (2.1%) 3 (2.1%) 31 (22.0%) 17 (12.1%) 10 (7.1%) 11 (7.8%) 48 (34.0%) 47 (33.3%) 27 (19.1%) 10 (7.1%) 2.9% 4.8% 10.7%
218 (39.9%) 6 (1.1%) 23 (4.2%) 115 (21.1%) 79 (14.5%) 57 (10.4%) 48 (8.8%) 233 (42.7%) 235 (43.0%) 134 (24.5%) 70 (12.8%) 4.6% 7.7% 12.4%
5 (6.6%) 2 (2.6%) 10 (13.2%) 6 (7.9%) 34 (44.7%) 7 (9.2%) 12 (15.8%) 42 (55.3%) 44 (57.9%) 34 (44.7%) 10 (13.2%) 12.8% 14.7% 17.4%
P value
<0.001 <0.001 <0.001 0.011 0.262 <0.001 <0.001
0.010 0.002 <0.001 0.158 0.001
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) or Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned (MDRD) formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker. Data are only shown for patients with eGFR 90 by either CKD-EPI or MDRD formulae.
294
Table 3 Characteristics at recruitment and outcome of patients reclassified using the CKD-EPI formula compared to the Lund-Malmö formula. 90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
15e29
<15
90
90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
<15
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem Venous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
324 51.6 (43.9e57.8) 163 (50.3%) 118 (36.4%) 55 (17.0%) 181 (55.9%) 53 (16.4%)
363 62.6 (56.7e67.3) 114 (31.4%) 194 (53.4%) 69 (19.0%) 262 (72.2%) 105 (28.9%)
785 69.2 (63.4e74.6) 225 (28.7%) 559 (71.2%) 193 (24.6%) 600 (76.4%) 339 (43.2%)
146 76.5 (71.9e81.1) 39 (26.7%) 109 (74.7%) 30 (20.5%) 110 (75.3%) 64 (43.8%)
238 74.2 (67.9e78.3) 73 (30.7%) 196 (82.4%) 64 (26.9%) 187 (78.6%) 124 (52.1%)
47 78.9 (75.5e82.3) 14 (29.8%) 44 (93.6%) 13 (27.7%) 31 (66.0%) 20 (42.6%)
128 76.3 (72.2e81.3) 38 (29.7%) 114 (89.1%) 45 (35.2%) 98 (76.6%) 80 (62.5%)
33 74.9 (71.2e79.6) 6 (18.2%) 26 (78.8%) 8 (24.2%) 30 (90.9%) 21 (63.6%)
54 75.6 (70.4e82.7) 22 (40.7%) 49 (90.7%) 26 (48.1%) 39 (72.2%) 37 (68.5%)
19 69.7 (55.7e77.7) 9 (47.4%) 19 (100%) 10 (52.6%) 13 (68.4%) 12 (63.2%)
178 (54.9%) 6 (1.9%)
106 (29.2%) 3 (0.8%)
116 (14.8%) 8 (1%)
8 (5.5%) 1 (0.7%)
8 (3.4%) 5 (2.1%)
1 (2.1%) 0 (0%)
1 (0.8%) 1 (0.8%)
1 (3.0%) 0 (0%)
1 (1.9%) 1 (1.9%)
0 (0%) 3 (15.8%)
6 (1.9%)
20 (5.5%)
61 (7.8%)
12 (8.2%)
29 (12.2%)
6 (12.8%)
9 (7.0%)
1 (3.0%)
4 (7.4%)
0 (0%)
63 (19.4%)
83 (22.9%)
127 (16.2%)
22 (15.1%)
41 (17.2%)
2 (4.3%)
13 (10.2%)
3 (9.1%)
8 (14.8%)
3 (15.8%)
21 (6.5%)
75 (20.7%)
286 (36.4%)
70 (47.9%)
107 (45%)
24 (51.1%)
78 (60.9%)
23 (69.7%)
25 (46.3%)
3 (15.8%)
25 (7.7%)
42 (11.6%)
105 (13.4%)
20 (13.7%)
27 (11.3%)
9 (19.1%)
10 (7.8%)
4 (12.1%)
2 (3.7%)
1 (5.3%)
25 (7.7%)
34 (9.4%)
82 (10.4%)
13 (8.9%)
21 (8.8%)
5 (10.6%)
16 (12.5%)
1 (3.0%)
13 (24.1%)
9 (47.4%)
97 (29.9%) 98 (30.2%) 67 (20.7%) 26 (8.0%) 6.0%
184 (50.7%) 184 (50.7%) 94 (25.9%) 54 (14.9%) 3.0%
463 (59.0%) 467 (59.5%) 287 (36.6%) 171 (21.8%) 7.6%
99 (67.8%) 105 (71.9%) 53 (36.3%) 35 (24.0%) 11.5%
154 (64.7%) 166 (69.7%) 100 (42.0%) 64 (26.9%) 11.2%
32 (68.1%) 33 (70.2%) 29 (61.7%) 10 (21.3%) 11.5%
82 (64.1%) 85 (66.4%) 62 (48.4%) 38 (29.7%) 14.8%
18 (54.5%) 16 (48.5%) 19 (57.6%) 8 (24.2%) 7.3%
36 (66.7%) 34 63.0%) 23 (42.6%) 18 (33.3%) 15.1%
13 (68.4%) 12 (63.2%) 8 (42.1%) 4 (21.1%) 41.7%
7.9%
6.6%
11.6%
21.6%
17.9%
25.2%
25.9%
11.3%
29.6%
61.2%
9.1%
13.7%
18.0%
27.8%
26.7%
39.7%
32.8%
37.8%
40.7%
61.2%
P value
<0.001 <0.001 <0.001 <0.001 0.262 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) or Lund-Malmö formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker.
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
eGFR range Lund-Malmö (ml/min/1.73 m2) eGFR range CKD-EPI (ml/min/1.73 m2)
Table 4 Characteristics at recruitment and outcome of patients reclassified using the Lund-Malmö formula compared to the MDRD formula. 90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
15e29
<15
90
90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
<15
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem enous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
319 51.9 (43.9e57.9) 161 (50.5%) 118 (37.0%) 54 (16.9%) 178 (55.8%) 53 (16.6%)
303 68.2 (62.6e73.9) 240 (79.2%) 203 (67.0%) 74 (24.4%) 234 (77.2%) 117 (38.6%)
839 66.5 (59.7e72.5) 567 (67.6%) 548 (65.3%) 187 (22.3%) 622 (74.1%) 326 (38.9%)
150 78.0 (73.7e82.7) 114 (76.0%) 111 (74.0%) 27 (18.0%) 116 (77.3%) 66 (44.0%)
233 73.0 (67.3e77.2) 158 (67.8%) 193 (82.8%) 67 (28.8%) 181 (77.7%) 121 (51.9%)
58 80.1 (75.9e82.8) 44 (75.9%) 53 (91.4%) 14 (24.1%) 41 (70.7%) 27 (46.6%)
117 75.4 (71.4e80.5) 79 (67.5%) 105 (89.7%) 44 (37.6%) 88 (75.2%) 73 (62.4%)
38 77.1 (72.2e85.6) 31 (81.6%) 29 (76.3%) 8 (21.1%) 31 (81.6%) 22 (57.9%)
49 74.6 (68.7e82.3) 28 (57.1%) 46 (93.9%) 26 (53.1%) 38 (77.6%) 36 (73.5%)
17 65.5 (54.9e77.5) 10 (58.8%) 17 (100%) 10 (58.8%) 11 (64.7%) 10 (58.8%)
173 (54.2%) 6 (1.9%)
50 (16.5%) 2 (0.7%)
170 (20.3%) 9 (1.1%)
6 (4.0%) 1 (0.7%)
10 (4.3%) 4 (1.7%)
1 (1.7%) 0 (0.0%)
1 (0.9%) 1 (0.9%)
1 (2.6%) 0 (0.0%)
1 (2.0%) 1 (2.0%)
0 (0.0%) 3 (17.6%)
6 (1.9%)
27 (8.9%)
53 (6.3%)
11 (7.3%)
30 (12.9%)
6 (10.3%)
9 (7.7%)
1 (2.6%)
4 (8.2%)
0 (0.0%)
63 (19.7%)
58 (19.1%)
151 (18.0%)
23 (15.3%)
40 (17.2%)
2 (3.4%)
13 (11.1%)
4 (10.5%)
7 (14.3%)
2 (11.8%)
21 (6.6%)
92 (30.4%)
268 (31.9%)
74 (49.3%)
103 (44.2%)
34 (58.6%)
68 (58.1%)
27 (71.1%)
21 (42.9%)
2 (11.8%)
25 (7.8%)
39 (12.9%)
107 (12.8%)
21 (14.0%)
26 (11.2%)
9 (15.5%)
10 (8.5%)
4 (10.5%)
2 (4.1%)
1 (5.9%)
25 (7.8%)
35 (11.6%)
81 (9.7%)
14 (9.3%)
20 (8.6%)
6 (10.3%)
15 (12.8%)
1 (2.6%)
13 (26.5%)
9 (52.9%)
96 (30.1%) 98 (30.7%) 67 (21.0%) 26 (8.2%) 6%
179 (59.1%) 181 (59.7%) 101 (33.3%) 54 (17.8) 5.9%
465 (55.4%) 468 (55.8%) 278 (33.1%) 171 (20.4%) 6.4%
96 (64.0%) 107 (71.3%) 52 (34.7%) 34 (22.7%) 10.9%
156 (67.0%) 164 (70.4%) 100 (42.9%) 65 (27.9%) 11.2%
36 (62.1%) 40 (69.0%) 33 (56.9%) 13 (22.4%) 15%
78 (66.7%) 78 (66.7%) 58 (49.6%) 35 (29.9%) 13.2%
22 (57.9%) 22 (57.9%) 21 (55.3%) 10 (26.3%) 6.6%
32 (65.3%) 28 (57.1%) 21 (42.9%) 16 (32.7%) 16.3%
12 (70.6%) 10 (58.8%) 7 (41.2%) 3 (17.6%) 40.1%
8%
9.5%
10.4%
21.6%
17.4%
28.6%
24.1%
10.5%
31.2%
64.1%
9.2%
15.6%
17.2%
29.0%
25.7%
40.9%
31.8%
35.7%
42.6%
64.1%
P value
<0.001 <0.001 <0.001 <0.001 <0.001 1.9% <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
eGFR range Lund-Malmö (ml/min/1.73 m2) eGFR range MDRD (ml/min/1.73 m2)
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Lund-Malmö or Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned (MDRD) formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker. Data for the 12 patients that were reclassified to a less advanced eGFR category (5 to eGFR 90; 6 to eGFR 60e89; 1 to eGFR 45e59 ml/min/1.73 m2) by the Lund-Malmö formula in comparison to the MDRD equation are not shown.
295
296
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
failure [36,37]. These previous studies, including a large metaanalysis of over one million adults, have reported that eGFR categorised by the CKD-EPI formula is more accurate at predicting adverse events than eGFR categorised by the MDRD formula [36,37]. Previous investigations that have assessed the association of eGFR with outcome in PVD patients however have almost always employed the MDRD formula [12e15,17e20]. We identified only one study that used the CKD-EPI formula to calculate eGFR as part of assessing the association of CKD with community outcomes in patients with PVD [11]. In that investigation eGFR was associated with outcome in 383 patients after undergoing endovascular aortic aneurysm repair. The current study suggests that research investigators and clinicians calculating eGFR as part of risk prediction in patients with PVD should consider using either the CKD-EPI or Lund-Malmö formulae. We identified no previous studies assessing the ability of eGFR categorised by the Lund-Malmö formula to predict major adverse events in any patient groups. This formula is not commonly used to assess eGFR however a recent study suggested that it gave the most accurate prediction of actual GFR in patients with late stage CKD [21]. In the current study eGFR categorised by the Lund-Malmö formula was equally as good at predicting major adverse events as eGFR categorised by the CKD-EPI formula. eGFR categorised by the Lund-Malmö formula was significantly better at predicting major adverse events than eGFR categorised by the MDRD formula. The Lund-Malmö formula defines substantially more patients to have CKD than both the CKD-EPI and MDRD formulae. Approximately an extra 15% of patients were categorised as having an eGFR <90 ml/ min/1.73 m2 and whether this is appropriate is controversial [10]. Overall our findings suggest both the CKD-EPI and Lund-Malmö formulae are appropriate to use in assessing risk of major adverse events in PVD patients. Other considerations apart from the accuracy of predicting major adverse events should also be considered before deciding on which formula to employ. These may include ability to compare with other studies, ease of calculating eGFR, availability of pre-calculated eGFR from local pathology departments and ability to predict other events such as progression to end stage renal failure. We did not examine these factors in the current study. A review of previous publications however suggests that the CKD-EPI formula is in more common use than the LundMalmö formula and many CKD foundations, such as Kidney Health Australia, recommend use of the CKD-EPI formula [38]. Taken together these factors favour the current use of the CKD-EPI formula in predicting eGFR in PVD patients. Our findings confirmed those from previous mainly smaller studies demonstrating an important increased risk of major adverse events in patients with PVD that have CKD reflected by reduced eGFR [12e20]. In the current study outcome was particularly poor for PVD patients with eGFR <45 with the estimated increased risk for patients categorised as eGFR 30e44, 15e29 and <15 ml/min/1.73 m2 approximately 1.5, 2 and 3-fold, respectively. The current study has a number of strengths and limitations. The sample size was relatively large compared to studies that have previously examined the association of CKD with outcome in PVD patients, although for some subgroups, such as those with eGFR <15 ml/min/1.73 m2 the numbers were small. The patients included in the current study presented with a range of different PVDs with varying severities of problems making the group representative of the overall population of patients with PVD but also quite heterogeneous. Thus our findings may have less applicability to very homogenous groups of patients with PVD such as those with lower limb athero-thrombosis alone. We examined three eGFR formulae although a number of others are available, such as the Cockroft-Gault, thus we can make no comment regarding the ability of other formulae to predict major adverse
events. We assessed eGFR only using serum creatinine and did not use other methods of assessing renal function such as serum cystatin C. Furthermore, proteinuria was not examined within the current study. Finally we did not assess the influence of amputation on eGFR. In conclusion the current study suggests that eGFR categorised by the CKD-EPI and Lund-Malmö formulae are equally effective at predicting major adverse events in PVD patients and their use is preferable to categorisation by the MDRD formula. Support and financial disclosure declaration Funding from the Queensland Government, the Townsville Hospital Private Practice Fund and National Health and Medical Research Council supported this work. JG holds a Practitioner Fellowships from the National Health and Medical Research Council, Australia (1019921). JG holds a Senior Clinical Research Fellowship from the Office of Health and Medical Research, Queensland. The authors have no conflict of interests relevant to this article. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atherosclerosis.2013.11.034. References [1] Eckardt KU, Coresh J, Devuyst O, et al. Evolving importance of kidney disease: from subspecialty to global health burden. Lancet 2013;382:158e69. [2] Gansevoort RT, Correa-Rotter R, Hemmelgarn BR, et al. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. Lancet 2013;382:339e52. [3] Layden J, Michaels J, Bermingham S, Higgins B, Guideline Development Group. Diagnosis and management of lower limb peripheral arterial disease: summary of NICE guidance. BMJ 2012;345:e4947. [4] Rooke TW, Hirsch AT, Misra S, et al. 2011 ACCF/AHA focused update of the guideline for the management of patients with peripheral artery disease (updating the 2005 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2011;58:2020e45. [5] European Stroke Organisation, Tendera M, Aboyans V, Bartelink ML, et al., ESC Committee for Practice Guidelines. ESC guidelines on the diagnosis and treatment of peripheral artery diseases: document covering atherosclerotic disease of extracranial carotid and vertebral, mesenteric, renal, upper and lower extremity arteries: the Task Force on the diagnosis and treatment of peripheral artery diseases of the European Society of Cardiology (ESC). Eur Heart J 2011;32:2851e906. [6] Nicolaides AN, Allegra C, Bergan J, et al. Management of chronic venous disorders of the lower limbs: guidelines according to scientific evidence. Int Angiol 2008;27:1e59. [7] Wattanakit K, Folsom AR, Selvin E, Coresh J, Hirsch AT, Weatherley BD. Kidney function and risk of peripheral arterial disease: results from the Atherosclerosis Risk in Communities (ARIC) Study. J Am Soc Nephrol 2007;18:629e36. [8] Garimella PS, Hart PD, O’Hare A, DeLoach S, Herzog CA, Hirsch AT. Peripheral artery disease and CKD: a focus on peripheral artery disease as a critical component of CKD care. Am J Kidney Dis 2012;60:641e54. [9] Florkowski CM, Chew-Harris JS. Methods of estimating GFRedifferent equations including CKD-EPI. Clin Biochem Rev 2011;32:75e9. [10] Moynihan R, Glassock R, Doust J. Chronic kidney disease controversy: how expanding definitions are unnecessarily labelling many people as diseased. BMJ 2013;347:f4298. [11] Saratzis A, Sarafidis P, Melas N, Saratzis N, Kitas G. Impaired renal function is associated with mortality and morbidity after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2013;58:879e85. [12] Lacroix P, Aboyans V, Desormais I, et al., COPART Investigators. Chronic kidney disease and the short-term risk of mortality and amputation in patients hospitalized for peripheral artery disease. J Vasc Surg 2013;58:966e71. [13] Liang KW, Kuo HN, Lee WL, et al. Different mid-term prognostic predictors of major adverse events in diabetic and nondiabetic peripheral artery disease presenting with critical limb ischemia. Angiology 2013 Feb 6 [Epub ahead of print]. [14] Romero JM, Bover J, Fite J, Bellmunt S, Dilmé JF, Camacho M, et al. The modification of diet in renal disease 4-calculated glomerular filtration rate is a better prognostic factor of cardiovascular events than classical cardiovascular risk factors in patients with peripheral arterial disease. J Vasc Surg 2012;56: 1324e30.
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297 [15] Kumakura H, Kanai H, Aizaki M, et al. The influence of the obesity paradox and chronic kidney disease on long-term survival in a Japanese cohort with peripheral arterial disease. Vasc Med 2011;16:422e7. [16] Ohrlander T, Dencker M, Dias NV, Gottsäter A, Acosta S. Cardiovascular predictors for long-term mortality after EVAR for AAA. J Vasc Surg 2010;52: 110e7. [17] Jaffery Z, Greenbaum AB, Siddiqui MF, et al. Predictors of mortality in patients with lower extremity peripheral arterial disease: 5-year follow-up. J Interv Cardiol 2009;22:564e70. [18] Pasqualini L, Schillaci G, Pirro M, et al. Renal dysfunction predicts long-term mortality in patients with lower extremity arterial disease. J Intern Med 2007;262:668e77. [19] Owens CD, Ho KJ, Kim S, et al. Refinement of survival prediction in patients undergoing lower extremity bypass surgery: stratification by chronic kidney disease classification. J Vasc Surg 2007;45:944e52. [20] O’Hare AM, Bertenthal D, Shlipak MG, Sen S, Chren MM. Impact of renal insufficiency on mortality in advanced lower extremity peripheral arterial disease. J Am Soc Nephrol 2005;16:514e9. [21] Evans M, van Stralen KJ, Schön S, et al., On the behalf of the ERA-EDTA Registry and the Swedish Renal Registry collaboration. Glomerular filtration rateestimating equations for patients with advanced chronic kidney disease. Nephrol Dial Transplant 2013;28:2518e26. [22] Golledge J, Cronin O, Iyer V, Bradshaw B, Moxon JV, Cunningham MA. Body mass index is inversely associated with mortality in patients with peripheral vascular disease. Atherosclerosis 2013;229:549e55. [23] Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373e9. [24] Palamuthusingam D, Quigley F, Golledge J. Implications of the finding of no significant carotid stenosis based on data from a regional Australian vascular unit. Ann Vasc Surg 2011;25:1050e6. [25] Parr A, Buttner P, Shahzad A, Golledge J. Relation of infra-renal abdominal aortic calcific deposits and cardiovascular events in patients with peripheral artery disease. Am J Cardiol 2010;105:895e9. [26] Golledge J, Jayalath R, Oliver L, Parr A, Schurgers L, Clancy P. Relationship between CT anthropometric measurements, adipokines and abdominal aortic calcification. Atherosclerosis 2008;197:428e34. [27] Parr A, McCann M, Bradshaw B, Shahzad A, Buttner P, Golledge J. Thrombus volume is associated with cardiovascular events and aneurysm growth in patients who have abdominal aortic aneurysms. J Vasc Surg 2011;53:28e35.
297
[28] Magee R, Quigley F, McCann M, Buttner P, Golledge J. Growth and risk factors for expansion of dilated popliteal arteries. Eur J Vasc Endovasc Surg 2010;39: 606e11. [29] Porter JM, Moneta GL. Reporting standards in venous disease: an update. International consensus committee on chronic venous disease. J Vasc Surg 1995;21:635e45. [30] Eknoyan G, Hostetter T, Bakris GL, et al. Proteinuria and other markers of chronic kidney disease: a position statement of the national kidney foundation (NKF) and the national institute of diabetes and digestive and kidney diseases (NIDDK). Am J Kidney Dis 2003;42:617e22. [31] Van Campenhout A, Moran CS, Parr A, et al. Role of homocysteine in aortic calcification and osteogenic cell differentiation. Atherosclerosis 2009;202: 557e66. [32] Björk J, Grubb A, Sterner G, Nyman U. Revised equations for estimating glomerular filtration rate based on the Lund-Malmö Study cohort. Scand J Clin Lab Invest 2011;71:232e9. [33] Anderson JL, Adams CD, Antman EM, et al., 2011 Writing Group Members; ACCF/AHA Task Force Members. 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 guidelines for the management of patients with unstable Angina/Non-ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011;123:e426e579. [34] Wright L, Hill KM, Bernhardt J, et al., National Stroke Foundation Stroke Guidelines Expert Working Group. Stroke management: updated recommendations for treatment along the care continuum. Intern Med J 2012;42: 562e9. [35] Pencina MJ, D’Agostino Sr RB, D’Agostino Jr RB, Vasan RS. Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond. Stat Med 2008;27:157e72. [36] Matsushita K, Mahmoodi BK, Woodward M, et al., Chronic Kidney Disease Prognosis Consortium. Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. J Am Med Assoc 2012;307:1941e51. [37] Oh J, Kang SM, Hong N, et al., KorHF Registry. The CKD-EPI is more accurate in clinical outcome prediction than MDRD equation in acute heart failure: data from the Korean Heart Failure (KorHF) registry. Int J Cardiol 2013;167:1084e7. [38] Kidney Health Australia. GFR calculator. Available from: http://www.kidney. org.au/HealthProfessionals/GFRCalculatorCKDEPI/tabid/803/Default.aspx. [accessed 30.08.13].
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Association of chronic kidney disease categories defined with different formulae with major adverse events in patients with peripheral vascular disease Jonathan Golledge a, b, *, Carla Ewels a, c, Reinhold Muller a, c, Phillip J. Walker d, e a
Queensland Research Centre for Peripheral Vascular Disease, School of Medicine and Dentistry, James Cook University, Townsville, Australia Department of Vascular and Endovascular Surgery, The Townsville Hospital, Townsville, Australia School of Public Health, Tropical Medicine and Rehabilitation Sciences, James Cook University, Townsville, Australia d University of Queensland School of Medicine, Discipline of Surgery and Centre for Clinical Research, Brisbane, Australia e Department of Vascular Surgery, The Royal Brisbane and Women’s Hospital, Brisbane, Australia b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 September 2013 Received in revised form 1 November 2013 Accepted 1 November 2013 Available online 2 December 2013
Objective: The aim of this study was to compare the ability of eGFR calculated by modification of diet in renal disease (MDRD), Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) and Lund-Malmö formulae in predicting major adverse events in peripheral vascular disease (PVD) patients. Methods: We prospectively recruited 2137 patients, measured serum creatinine to calculate eGFR using three different formulae and grouped patients into eGFR categories 90, 60e89, 45e59, 30e44, 15e29 and <15 ml/min/1.73 m2. Patients were followed up for a median of 1.3 (inter-quartile range 0.3e3.6) years. The primary outcome was the combined incidence of myocardial infarction, stroke or death. The ability of eGFR categories defined with the different formulae to predict outcome was assessed using the net reclassification index. Results: 1450 (67.9%), 1515 (70.9%) and 1813 (84.8%) patients had eGFR <90 ml/min/1.73 m2 according to the CKD-EPI, MDRD and Lund-Malmö formulae, respectively. Using the CKD-EPI formula 276 (12.9%) patients were reclassified to a different eGFR category in comparison to the MDRD formula and the prediction of outcome was improved (net reclassification index 0.106, p < 0.001). Using the Lund-Malmö formula 563 (26.3%) patients were reclassified to a different eGFR category in comparison to the MDRD formula and the prediction of outcome was improved (net reclassification index 0.108, p < 0.001). Classification using the CKD-EPI and Lund-Malmö formulae was equally effective at predicting outcome (net reclassification index - 0.002, p ¼ 0.397). Conclusions: eGFR categories determined with the CKD-EPI and Lund-Malmö formulae are equally effective at predicting major adverse events in patients with PVD. Ó 2013 Elsevier Ireland Ltd. All rights reserved.
Keywords: Peripheral vascular disease Chronic kidney disease Estimated glomerular filtration rate
1. Introduction The prevalence of chronic kidney disease (CKD) is estimated to be >10% and increasing [1]. The prevalence of CKD is particularly high in some patients groups, such as patients with cardiovascular disease [2]. Peripheral vascular diseases (PVD) are a group of conditions affecting the vessels outside the heart [3e6]. CKD is a risk factor for PVD development and CKD patients with PVD have an increased incidence of major cardiovascular events [7,8]. There is
* Correspondence author. Queensland Research Centre for Peripheral Vascular Disease, School of Medicine and Dentistry, James Cook University, Townsville 4811, Australia. Tel.: þ61 7 4796 1417; fax þ61 7 4796 1401. E-mail address: [email protected] (J. Golledge). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.11.034
current controversy over how CKD is best defined with a large number of different ways to estimate kidney function by calculating estimated glomerular filtrate rate (eGFR) being available [9,10]. A number of previous studies have associated CKD with major adverse events in patients with PVD [11e20]. In these studies a number of different equations have been used to calculate eGFR, including the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) and the modification of diet in renal disease (MDRD) formulae [11e20]. Recently eGFR calculated using the Lund-Malmö formula has been suggested as being more accurate in some patients groups [21]. There are a number of possible reasons to suspect that the most appropriate formula to assess eGFR in PVD patients might be different from healthy individuals. The prevalence of CKD is high amongst PVD patients and eGFR formulae vary in their ability to estimate severe renal function impairment
290
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
[1,2,9,10]. Furthermore, the aetiology of CKD in PVD is more likely to be related to renovascular disease which may also influence the most appropriate formula to use [1,2,9,10]. Currently it is not clear which formula best predicts major adverse events in patients with PVD. The primary aim of this study was to assess which of three formulae used to calculate eGFR best predicted major adverse events in patients with PVD. 2. Methods 2.1. Study design This study was designed as an on-going prospective cohort investigation of patients with PVD aimed at assessing risk predictors of PVD presence and outcome commencing in 2002, as previously described [22]. Since the comparative ability of different eGFR formulae to predict major adverse events in PVD patients had not been previously examined sample size calculations were not straight forward and no formal calculation was performed [11e20]. Monte-Carlo simulations suggest that a multivariate regression model is powered sufficiently when 10 outcome events per degree of freedom of the predictor variables are observed [23]. We estimated that the combined incidence of myocardial infarction, stroke or death at one year would be approximately 10% and planned to adjust for 11 variables (age, male sex, hypertension, diabetes, smoking, coronary heart disease (CHD), presenting complaint, statin prescription, aspirin prescription, angiotensin converting enzyme (ACE) inhibitor prescription and angiotensin receptor blocker (ARB) prescription) in our regression model. Based on these estimates we felt that a sample size of over 2000 patients would be well powered to examine the association of eGFR categories with major adverse events. 2.2. Patients
This included telangiectasia or reticular veins (C1); varicose veins of 3 mm (C2); oedema (C3); skin changes due to chronic venous disease (C4); healed venous ulcer (C5); and active venous ulcer (C6). All patients underwent venous duplex imaging; b) Miscellaneous PVD problems including aortic dissection, reno-vascular hypertension, mesenteric ischaemia and peripheral vascular trauma were diagnosed based on history, examination and imaging using duplex imaging or computed tomographic angiography; c) Asymptomatic carotid artery stenosis: Defined as the presence of 50% stenosis or occlusion of at least one carotid artery identified by carotid duplex but the absence of physician confirmed symptoms of focal transient ischaemic attack, amaurosis fugax or stroke as previously described [24]; d) Mild lower limb or upper limb peripheral athero-thrombosis: This included patients with intermittent claudication, atypical or no symptoms with clinical evidence of lower or upper limb ischaemia but not critical lower limb ischaemia. Limb peripheral athero-thrombosis was confirmed by a vascular specialist by identification of absence of lower or upper limb pulses, ankle brachial pressure index <0.9 and/or significant stenosis (>50%) or occlusion of lower or upper limb arteries on computed tomographic angiography or duplex imaging [25,26]. e) Aneurysm of the aorta or peripheral arteries: Aortic aneurysm was defined as maximum aortic diameter 30 mm [25e27]. Iliac artery aneurysm was defined by common or internal iliac artery diameters 15 and 8 mm, respectively. Femoral artery aneurysm was defined by common femoral or superficial femoral artery diameter of 15 mm. Popliteal artery aneurysm was defined as popliteal artery diameter 9 mm as previously described [28]; f) Symptomatic carotid artery stenosis: Defined as the presence of 50% stenosis or occlusion of at least one carotid artery identified with carotid duplex with the presence of physician confirmed symptoms of focal transient ischaemic attack, amaurosis fugax or stroke as previously described [24]; g) Critical lower limb ischaemia: Rest pain, arterial ulcer or gangrene of the leg due to athero-thrombosis of the lower limb. Peripheral athero-thrombosis was confirmed as detailed above [25,26]. For patients with more than one presenting complaint classification was determined by the complaint which was deemed most severe.
Patients were recruited from in and out-patient vascular services at The Townsville Hospital, The Mater Hospital Townsville and The Royal Brisbane and Women’s Hospital. Patients with all types of PVD were considered for inclusion. All patients diagnosed as having any type of PVD by a Royal Australasian College of Surgeons accredited vascular specialist were considered for inclusion into the study. Inclusion criteria for the current study included a diagnosis of PVD, the assessment of serum creatinine to enable the calculation of eGFR and a least one follow-up assessment as an in or outpatient. Ethical approval for the study was granted by the local Institutional Ethics Committees at The Townsville Hospital, The Mater Hospital Townsville, The Royal Brisbane and Women’s Hospital and James Cook University. Written informed consent was obtained from participants.
Hypertension was defined by a history of high blood pressure or receiving treatment to reduce blood pressure [22,24e28]. Diabetes was defined by a fasting blood glucose concentration 7.0 mM, or history of, or treatment for hyperglycaemia [22,24e28]. Smoking status was classified as ever and never smokers [22,24e28]. CHD was defined by a history of myocardial infarction, angina or treatment for coronary artery disease [22,24e28].
2.3. Definition of presenting complaint
2.6. Medications
Presenting category was broadly defined into one of seven groups namely venous disease; miscellaneous PVDs (including aortic dissection, reno-vascular hypertension, mesenteric ischaemia and peripheral vascular trauma); asymptomatic carotid stenosis; mild lower limb or upper limb peripheral atherothrombosis; aneurysm of the aorta or peripheral arteries; symptomatic carotid artery stenosis; and critical lower limb ischaemia, as previously described in detail [22,24e28].
At the time of recruitment a list of each patient’s medications was recorded including whether the participants were prescribed statins, aspirin, ACE inhibitors or ARBs.
2.4. Definitions and diagnosis of PVD PVD was defined using the following criteria: a) Venous disease: This was defined according to the CEAP classification [29].
2.5. Definition of other risk factors
2.7. Measurement of serum creatinine and calculation of eGFR Serum creatinine was measured in a pathology laboratory using a spectrophotometry method in line with established guidelines as previously described [30,31]. eGFR was calculated using the CKD-EPI, MDRD (isotope dilution mass spectrometry aligned) and Lund-Malmö formulae [9,21,32]. All these formulae have been previously described in detail and utilise creatinine, age and gender in their calculations [9,21,32]. The following
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
291
eGFR categories were defined: 90, 60e89, 45e59, 30e44, 15e 29 and <15 ml/min/1.73 m2 as currently used to define CKD stages [1].
3. Results
2.8. Follow-up
The presenting complaints of the 2137 patients included aortic or peripheral aneurysms (n ¼ 712; 33.3%), venous disease (n ¼ 420; 19.7%), non-critical limb athero-thrombosis (n ¼ 365; 17.1%), symptomatic carotid artery stenosis (n ¼ 245; 11.5%), critical lower limb ischaemia (n ¼ 219; 10.2%), asymptomatic carotid artery stenosis (n ¼ 148; 6.9%), and miscellaneous PVDs (n ¼ 28; 1.3%). Sixty seven per cent of patients were male, 73% had a history of ever smoking, 67% had a history of hypertension, 40% a history of CHD and 24% had diabetes (Supplementary Table 1). Approximately half the patients were receiving a statin and aspirin, 35% receiving an angiotensin converting enzyme inhibitor and 20% receiving an angiotensin receptor blocker.
Patients were followed up through attendance at out-patient clinics and/or as an in-patient as part of their normal medical care. Patients with venous disease and miscellaneous PVD were followed up as needed depending on the severity of their presenting complaint and any medical intervention planned. Patients with limb athero-thrombosis and carotid artery stenosis were generally reviewed 6 months after their initial assessment and then yearly unless symptoms or imaging findings changed [23e25]. Patients with small aneurysms or large aneurysms that had been repaired were followed up yearly or 6 monthly if the aneurysm was nearing a diameter at which intervention was indicated [26,27].
3.1. Characteristics of the cohort
3.2. Relationship between risk factors and eGFR categories
The primary outcome was the combined incidence of myocardial infarction, stroke or death. Charts and hospital electronic records of all patients were reviewed at least once for the identification and date of any outcome events by a vascular specialist or research nurse. Where uncertainty about outcome data was present a discussion occurred between the research nurse and vascular specialist and a consensus was reached. Myocardial infarctions and strokes were defined based on diagnosis by a Royal Australasian College of Physicians accredited physician in line with international guidelines [33,34]. For patients that had not experienced the primary outcome follow-up was concluded, i.e. censoring occurred, at the date of last in or outpatient review.
eGFR calculated by the CKD-EPI formula (median 80.3, interquartile range, IQR, 60.8e93.3 ml/min/1.73 m2) was significantly higher than that estimated by both the MDRD equation (median 76.6, IQR 60.6e93.6 ml/min/1.73 m2 p ¼ 0.014, Wilcoxon Signed Rank Test) and Lund-Malmö (median 70.3, IQR 55.9e83.0 ml/min/ 1.73 m2 p < 0.001, Wilcoxon Signed Rank Test) formulae. Nine patients had end-stage renal failure requiring dialysis and had eGFR <15 ml/min/1.73 m2. Overall patients with eGFR 90 ml/min/ 1.73 m2 defined by the CKD-EPI formula were more likely to be younger, female, never smokers and present with venous disease (Supplementary Table 1). Patients with eGFR 90 ml/min/1.73 m2 were less likely to have hypertension, diabetes, CHD, or be prescribed statins, aspirin, ACE inhibitors and angiotensin receptor blockers (Supplementary Table 1). Similar associations were noted when eGFR was defined using the MDRD and Lund-Malmö formulae (data not shown).
2.10. Statistical analyses
3.3. Relationship between outcome and eGFR
Quantitative data were not normally distributed and therefore are presented as median and inter-quartile range and assessed by non-parametric tests. Nominal data are presented as number and percentages and compared by chi-squared test. The associations of eGFR and eGFR categories with the combined incidence of myocardial infarction, stroke or death were assessed using Kaplan Meier estimates and Cox proportional hazard analyses. Cox proportional hazard analysis was adjusted for age >68 years (median age of the cohort), male sex, hypertension, diabetes, smoking, CHD, presenting complaint, statin prescription, aspirin prescription, ACE inhibitor prescription and ARB prescription. For the Cox analyses presenting PVD problem was defined as an indicator from one to seven in the following order: venous disease; miscellaneous PVDs; asymptomatic carotid stenosis; mild lower limb or upper limb peripheral athero-thrombosis; aneurysm of the aorta or peripheral arteries; symptomatic carotid artery stenosis; and critical lower limb ischaemia. Other adjusted variables were defined as present or absent. The variables adjusted for were chosen as they are recognised determinants of outcome for patients with cardiovascular disease. The combined incidence of myocardial infarction, stroke or death was compared between groups of subjects in different eGFR categories using the log rank test. The ability of eGFR categories defined with the different formulae to predict the combined incidence of myocardial infarction, stroke or death was assessed using the net reclassification index which was calculated as previously described [35].
Patients were followed up for a median of 1.3 (inter-quartile range 0.3e3.6) years. During this period 106 patients had a myocardial infarction, 72 patients had a stroke and 327 patients died. 42 of the patients who suffered a myocardial infarction and 33 of those who experienced a stroke subsequently died during later follow-up. The primary outcome of myocardial infarction, stroke or death occurred in 425 independent patients while the other 1712 patients were censored at last follow-up. Overall the estimated combined incidences of myocardial infarction, stroke or death were 8.6, 14.2 and 21.5% at 1, 2 and 3 years, respectively. The estimated combined incidence of myocardial infarction, stroke or death at three years were 12.1, 19.6, 29.7, 34.7, 40.7 and 61.2% in patients that had an CKD-EPI eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1a). The estimated combined incidence of myocardial infarction, stroke or death at three years were 13.3, 19.1, 29.2, 34.0, 42.8 and 64.1% in patients that had an MDRD eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1b). The estimated combined incidence of myocardial infarction, stroke or death at three years were 9.1, 16.7, 27.2, 35.8, 39.7 and 61.2% in patients that had an Lund-Malmö eGFR of 90, 60e89; 45e59; 30e44; 15e29 and <15 ml/min/1.73 m2 at recruitment, respectively, p < 0.001 (Fig. 1c). After adjustment for age, male gender, hypertension, diabetes, ever smoking, CHD, presenting problem and prescription of statins, aspirin, ACE inhibitors and ARBs a 10 ml/min/1.73 m2 increase in eGFR (approximately one standard deviation in the whole population) was
2.9. Recording of outcome data
292
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
Fig. 1. Kaplan Meier curves showing the incidence of myocardial infarction, stroke or death in relation to eGFR category at entry defined with the a) CKD-EPI; b) MDRD; and c) LundMalmö formulae. Vertical lines represent subjects censored at loss to follow-up.
associated with a reduced incidence of myocardial infarction, stroke or death although this finding was stronger for values estimated with the CKD-EPI and Lund-Malmö formulae (CKD-EPI RR 0.90, 95% CI 0.86e0.94, p < 0.001; MDRD RR 0.96, 95% CI 0.92e1.00, p ¼ 0.027; Lund-Malmö RR 0.88, 95% CI 0.84e0.93, p < 0.001). eGFR categories 30e44, 15e29 and <15 ml/min/1.73 m2 were independently associated with increased incidence of myocardial infarction, stroke or death when calculated using the CKD-EPI and MDRD formulae (Table 1). Only eGFR category <15 ml/min/1.73 m2 was independently associated with increased incidence of myocardial infarction, stroke or death when calculated using the Lund-Malmö formula (Table 1). 3.4. eGRF categories according to the different formulae and relationship with risk factors and outcomes 1450 (67.9%), 1515 (70.9%) and 1813 (84.8%) patients had eGFR <90 ml/min/1.73 m2 according to the CKD-EPI, MDRD and LundMalmö formulae, respectively. Using the CKD-EPI formula 113 (5.3%) patients were reclassified to a worse eGFR category (76 to eGFR 60e89; 18 to eGFR 45e59; 11 to eGFR 30e44; 6 to eGFR 15e 29 and 2 to eGFR <15 ml/min/1.73 m2) and 163 (7.6%) to a less advanced eGFR category (141 to eGFR 90; 20 to eGFR 60e89; 1 to eGFR 45e59; and 1 to eGFR 30e44 ml/min/1.73 m2) in comparison to classification using the MDRD formula. Thus most of the
reclassification was for patients who were categorised to have an eGFR 90 or 60e89 ml/min/1.73 m2 according to the MDRF formula. 76 of 622 (12.2%) patients classified to have an eGFR 90 with the MDRF formula were reclassified to have an eGFR of 60e89 ml/ min/1.73 m2 using the CKD-EPI formula. 141 of 994 (14.2%) patients classified to have an eGFR of 60e89 with the MDRF formula were reclassified to have an eGFR 90 ml/min/1.73 m2 using the CKD-EPI formula. The risk factors and outcomes of patients who had an eGFR 90 according to one or both formulae in relation to whether they were reclassified by the CKD-EPI formula are shown in Table 2. Patients reclassified to a worse eGFR category were older and more likely to be male, have hypertension, have diabetes, have CHD and present with an aneurysm or critical limb ischaemia. They were also more likely to be prescribed a statin, aspirin and an ACE inhibitor. Patients reclassified to a worse category had increased incidence of myocardial infarction, stroke or death while patients reclassified to a better eGFR category had reduced incidence of these events (Table 2, Supplementary Fig. 1). The net reclassification index improvement using the CKD-EPI as opposed to the MDRD formula was 0.106, p < 0.001. Using the CKD-EPI formula no patients were reclassified to a worse eGFR category but 589 (27.6%) were reclassified to a less advanced eGFR category (363 to eGFR 90; 146 to eGFR 60e89; 47 to eGFR 45e59; and 33 to eGFR 30e44 ml/min/1.73 m2) in comparison to classification using the Lund-Malmö formula. Generally
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297 Table 1 Association of eGFR categories calculated with CKD-EPI, MDRD and Lund-Malmö formulae with major adverse events (myocardial infarction, stroke or death) in 2137 patients with peripheral vascular disease eGFR category (ml/min/1.73 m2 )
Relative risk
Calculated with CKD-EPI formula 90 1.00 60e89 0.98 45e59 1.23 30e44 1.59 15e29 1.90 <15 3.38 Calculated with MDRD formula 90 1.00 60e89 0.97 45e59 1.21 30e44 1.48 15e29 2.06 <15 3.17 Calculated with Lund-Malmö formula 90 1.00 60e89 0.88 45e59 1.14 30e44 1.42 15e29 1.66 <15 3.14
95% CI
P value
Reference group 0.72e1.33 0.86e1.75 1.07e2.36 1.14e3.16 1.68e6.80
0.903 0.250 0.022 0.014 0.001
Reference group 0.74e1.27 0.88e1.66 1.02e2.15 1.28e3.29 1.49e6.78
0.843 0.249 0.037 0.003 <0.001
Reference group 0.56e1.39 0.69e1.86 0.84e2.43 0.95e2.93 1.44e6.84
0.575 0.614 0.194 0.078 0.004
eGFR ¼ Estimated glomerular filtration; CKD-EPI ¼ Chronic Kidney DiseaseEpidemiology Collaboration group formula; MDRD ¼ Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned formula; All models were adjusted for age above 68 years, male gender, coronary heart disease, hypertension, diabetes, ever smoker, presenting problem, statin prescription, aspirin prescription, angiotensin converting enzyme inhibitor prescription and angiotensin receptor blocker prescription.
patients reclassified to less severe eGFR categories were less likely to have established cardiovascular risk factors than those classified in the same eGFR category according to both formulae (Table 3). The relationship between reclassification using the CKD-EPI and the Lund-Malmö formulae and incidence of myocardial infarction, stroke or death are shown in Table 3 and Supplementary Fig. 2.
293
Patients reclassified from the 60e89 to 90 ml/min/1.73 m2 category by the CKD-EPI formula had a better prognosis than those in which eGFR was estimated to be 60e89 ml/min/1.73 m2 using both formulae. For other reclassifications the relationship was less clear (Table 3). The net reclassification index using the CKD-EPI as opposed to the Lund-Malmö formula was 0.002, p ¼ 0.397. Using the Lund-Malmö formula 551 (25.8%) patients were reclassified to a worse eGFR category (303 to eGFR 60e89; 150 to eGFR 45e59; 58 to eGFR 30e44; 38 to eGFR 15e29 and 2 to eGFR <15 ml/min/1.73 m2) and 12 (0.6%) to a less advanced eGFR category (5 to eGFR 90; 6 to eGFR 60e89; 1 to eGFR 45e59 ml/min/ 1.73 m2) in comparison to classification using the MDRD formula. Generally patients reclassified to more severe eGFR categories were more likely to have established cardiovascular risk factors than those classified as the same eGFR category according to both formulae (Table 4). The relationship between reclassification using the Lund-Malmö and the MDRD formulae and incidence of myocardial infarction, stroke or death is shown in Table 4. Patients reclassified from the 90 and 60e89 ml/min/1.73 m2 to more severe eGFR categories by the Lund-Malmö formula had a worse prognosis than those in which eGFR was the same using both formulae. For other reclassifications the relationship was less clear although the numbers of patients were small (Table 4). The net reclassification index improvement using the Lund-Malmö formula as opposed to the MDRD formula was 0.108, p < 0.001. 4. Discussion The current study suggests that the CKD-EPI and Lund-Malmö formulae are equally effective at defining patients with PVD into eGFR categories that are predictive of major adverse events. In contrast eGFR categories defined with the MDRD formula are inferior at predicting major adverse events for PVD patients. Our findings are in keeping with reports which have compared the use of the CKD-EPI and MDRD formulae in other populations, such as healthy older adults and patients presenting with heart
Table 2 Characteristics at recruitment and outcome of patients reclassified using the CKD-EPI formula compared to the MDRD formula. eGFR range MDRD (ml/min/1.73 m2) eGFR range CKD-EPI (ml/min/1.73 m2)
60e89 90
90 90
90 60e89
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem Venous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
141 56.3 (47.8e61.7) 68 (48.2%) 54 (38.3%) 19 (13.5%) 85 (60.3%) 26 (18.4%)
546 58.6 (48.7e66.0) 342 (62.6%) 258 (47.3%) 105 (19.2%) 358 (65.6%) 132 (24.2%)
76 76.2 (73.1e78.9) 59 (77.6%) 63 (82.9%) 23 (30.3%) 54 (71.1%) 38 (50.0%)
66 (46.8%) 3 (2.1%) 3 (2.1%) 31 (22.0%) 17 (12.1%) 10 (7.1%) 11 (7.8%) 48 (34.0%) 47 (33.3%) 27 (19.1%) 10 (7.1%) 2.9% 4.8% 10.7%
218 (39.9%) 6 (1.1%) 23 (4.2%) 115 (21.1%) 79 (14.5%) 57 (10.4%) 48 (8.8%) 233 (42.7%) 235 (43.0%) 134 (24.5%) 70 (12.8%) 4.6% 7.7% 12.4%
5 (6.6%) 2 (2.6%) 10 (13.2%) 6 (7.9%) 34 (44.7%) 7 (9.2%) 12 (15.8%) 42 (55.3%) 44 (57.9%) 34 (44.7%) 10 (13.2%) 12.8% 14.7% 17.4%
P value
<0.001 <0.001 <0.001 0.011 0.262 <0.001 <0.001
0.010 0.002 <0.001 0.158 0.001
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) or Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned (MDRD) formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker. Data are only shown for patients with eGFR 90 by either CKD-EPI or MDRD formulae.
294
Table 3 Characteristics at recruitment and outcome of patients reclassified using the CKD-EPI formula compared to the Lund-Malmö formula. 90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
15e29
<15
90
90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
<15
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem Venous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
324 51.6 (43.9e57.8) 163 (50.3%) 118 (36.4%) 55 (17.0%) 181 (55.9%) 53 (16.4%)
363 62.6 (56.7e67.3) 114 (31.4%) 194 (53.4%) 69 (19.0%) 262 (72.2%) 105 (28.9%)
785 69.2 (63.4e74.6) 225 (28.7%) 559 (71.2%) 193 (24.6%) 600 (76.4%) 339 (43.2%)
146 76.5 (71.9e81.1) 39 (26.7%) 109 (74.7%) 30 (20.5%) 110 (75.3%) 64 (43.8%)
238 74.2 (67.9e78.3) 73 (30.7%) 196 (82.4%) 64 (26.9%) 187 (78.6%) 124 (52.1%)
47 78.9 (75.5e82.3) 14 (29.8%) 44 (93.6%) 13 (27.7%) 31 (66.0%) 20 (42.6%)
128 76.3 (72.2e81.3) 38 (29.7%) 114 (89.1%) 45 (35.2%) 98 (76.6%) 80 (62.5%)
33 74.9 (71.2e79.6) 6 (18.2%) 26 (78.8%) 8 (24.2%) 30 (90.9%) 21 (63.6%)
54 75.6 (70.4e82.7) 22 (40.7%) 49 (90.7%) 26 (48.1%) 39 (72.2%) 37 (68.5%)
19 69.7 (55.7e77.7) 9 (47.4%) 19 (100%) 10 (52.6%) 13 (68.4%) 12 (63.2%)
178 (54.9%) 6 (1.9%)
106 (29.2%) 3 (0.8%)
116 (14.8%) 8 (1%)
8 (5.5%) 1 (0.7%)
8 (3.4%) 5 (2.1%)
1 (2.1%) 0 (0%)
1 (0.8%) 1 (0.8%)
1 (3.0%) 0 (0%)
1 (1.9%) 1 (1.9%)
0 (0%) 3 (15.8%)
6 (1.9%)
20 (5.5%)
61 (7.8%)
12 (8.2%)
29 (12.2%)
6 (12.8%)
9 (7.0%)
1 (3.0%)
4 (7.4%)
0 (0%)
63 (19.4%)
83 (22.9%)
127 (16.2%)
22 (15.1%)
41 (17.2%)
2 (4.3%)
13 (10.2%)
3 (9.1%)
8 (14.8%)
3 (15.8%)
21 (6.5%)
75 (20.7%)
286 (36.4%)
70 (47.9%)
107 (45%)
24 (51.1%)
78 (60.9%)
23 (69.7%)
25 (46.3%)
3 (15.8%)
25 (7.7%)
42 (11.6%)
105 (13.4%)
20 (13.7%)
27 (11.3%)
9 (19.1%)
10 (7.8%)
4 (12.1%)
2 (3.7%)
1 (5.3%)
25 (7.7%)
34 (9.4%)
82 (10.4%)
13 (8.9%)
21 (8.8%)
5 (10.6%)
16 (12.5%)
1 (3.0%)
13 (24.1%)
9 (47.4%)
97 (29.9%) 98 (30.2%) 67 (20.7%) 26 (8.0%) 6.0%
184 (50.7%) 184 (50.7%) 94 (25.9%) 54 (14.9%) 3.0%
463 (59.0%) 467 (59.5%) 287 (36.6%) 171 (21.8%) 7.6%
99 (67.8%) 105 (71.9%) 53 (36.3%) 35 (24.0%) 11.5%
154 (64.7%) 166 (69.7%) 100 (42.0%) 64 (26.9%) 11.2%
32 (68.1%) 33 (70.2%) 29 (61.7%) 10 (21.3%) 11.5%
82 (64.1%) 85 (66.4%) 62 (48.4%) 38 (29.7%) 14.8%
18 (54.5%) 16 (48.5%) 19 (57.6%) 8 (24.2%) 7.3%
36 (66.7%) 34 63.0%) 23 (42.6%) 18 (33.3%) 15.1%
13 (68.4%) 12 (63.2%) 8 (42.1%) 4 (21.1%) 41.7%
7.9%
6.6%
11.6%
21.6%
17.9%
25.2%
25.9%
11.3%
29.6%
61.2%
9.1%
13.7%
18.0%
27.8%
26.7%
39.7%
32.8%
37.8%
40.7%
61.2%
P value
<0.001 <0.001 <0.001 <0.001 0.262 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Chronic Kidney Disease-Epidemiology Collaboration group (CKD-EPI) or Lund-Malmö formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker.
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
eGFR range Lund-Malmö (ml/min/1.73 m2) eGFR range CKD-EPI (ml/min/1.73 m2)
Table 4 Characteristics at recruitment and outcome of patients reclassified using the Lund-Malmö formula compared to the MDRD formula. 90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
15e29
<15
90
90
60e89
60e89
45e59
45e59
30e44
30e44
15e29
<15
Number Age (years) Male Hypertension Diabetes mellitus Ever smoker CHD Presenting problem enous disease Miscellaneous PVDs Asymptomatic carotid stenosis Mild lower or upper limb PAD Aortic and peripheral artery aneurysms Symptomatic carotid artery stenosis Critical limb ischaemia Statin Aspirin ACE inhibitor ARB Incidence of stroke, MI or death at 1 year Incidence of stroke, MI or death at 2 years Incidence of stroke, MI or death at 3 years
319 51.9 (43.9e57.9) 161 (50.5%) 118 (37.0%) 54 (16.9%) 178 (55.8%) 53 (16.6%)
303 68.2 (62.6e73.9) 240 (79.2%) 203 (67.0%) 74 (24.4%) 234 (77.2%) 117 (38.6%)
839 66.5 (59.7e72.5) 567 (67.6%) 548 (65.3%) 187 (22.3%) 622 (74.1%) 326 (38.9%)
150 78.0 (73.7e82.7) 114 (76.0%) 111 (74.0%) 27 (18.0%) 116 (77.3%) 66 (44.0%)
233 73.0 (67.3e77.2) 158 (67.8%) 193 (82.8%) 67 (28.8%) 181 (77.7%) 121 (51.9%)
58 80.1 (75.9e82.8) 44 (75.9%) 53 (91.4%) 14 (24.1%) 41 (70.7%) 27 (46.6%)
117 75.4 (71.4e80.5) 79 (67.5%) 105 (89.7%) 44 (37.6%) 88 (75.2%) 73 (62.4%)
38 77.1 (72.2e85.6) 31 (81.6%) 29 (76.3%) 8 (21.1%) 31 (81.6%) 22 (57.9%)
49 74.6 (68.7e82.3) 28 (57.1%) 46 (93.9%) 26 (53.1%) 38 (77.6%) 36 (73.5%)
17 65.5 (54.9e77.5) 10 (58.8%) 17 (100%) 10 (58.8%) 11 (64.7%) 10 (58.8%)
173 (54.2%) 6 (1.9%)
50 (16.5%) 2 (0.7%)
170 (20.3%) 9 (1.1%)
6 (4.0%) 1 (0.7%)
10 (4.3%) 4 (1.7%)
1 (1.7%) 0 (0.0%)
1 (0.9%) 1 (0.9%)
1 (2.6%) 0 (0.0%)
1 (2.0%) 1 (2.0%)
0 (0.0%) 3 (17.6%)
6 (1.9%)
27 (8.9%)
53 (6.3%)
11 (7.3%)
30 (12.9%)
6 (10.3%)
9 (7.7%)
1 (2.6%)
4 (8.2%)
0 (0.0%)
63 (19.7%)
58 (19.1%)
151 (18.0%)
23 (15.3%)
40 (17.2%)
2 (3.4%)
13 (11.1%)
4 (10.5%)
7 (14.3%)
2 (11.8%)
21 (6.6%)
92 (30.4%)
268 (31.9%)
74 (49.3%)
103 (44.2%)
34 (58.6%)
68 (58.1%)
27 (71.1%)
21 (42.9%)
2 (11.8%)
25 (7.8%)
39 (12.9%)
107 (12.8%)
21 (14.0%)
26 (11.2%)
9 (15.5%)
10 (8.5%)
4 (10.5%)
2 (4.1%)
1 (5.9%)
25 (7.8%)
35 (11.6%)
81 (9.7%)
14 (9.3%)
20 (8.6%)
6 (10.3%)
15 (12.8%)
1 (2.6%)
13 (26.5%)
9 (52.9%)
96 (30.1%) 98 (30.7%) 67 (21.0%) 26 (8.2%) 6%
179 (59.1%) 181 (59.7%) 101 (33.3%) 54 (17.8) 5.9%
465 (55.4%) 468 (55.8%) 278 (33.1%) 171 (20.4%) 6.4%
96 (64.0%) 107 (71.3%) 52 (34.7%) 34 (22.7%) 10.9%
156 (67.0%) 164 (70.4%) 100 (42.9%) 65 (27.9%) 11.2%
36 (62.1%) 40 (69.0%) 33 (56.9%) 13 (22.4%) 15%
78 (66.7%) 78 (66.7%) 58 (49.6%) 35 (29.9%) 13.2%
22 (57.9%) 22 (57.9%) 21 (55.3%) 10 (26.3%) 6.6%
32 (65.3%) 28 (57.1%) 21 (42.9%) 16 (32.7%) 16.3%
12 (70.6%) 10 (58.8%) 7 (41.2%) 3 (17.6%) 40.1%
8%
9.5%
10.4%
21.6%
17.4%
28.6%
24.1%
10.5%
31.2%
64.1%
9.2%
15.6%
17.2%
29.0%
25.7%
40.9%
31.8%
35.7%
42.6%
64.1%
P value
<0.001 <0.001 <0.001 <0.001 <0.001 1.9% <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
eGFR range Lund-Malmö (ml/min/1.73 m2) eGFR range MDRD (ml/min/1.73 m2)
Nominal variables are presented as numbers (%) and compared by chi-squared. Continuous variables are presented as median (inter-quartile range) and compared by Kruskal Wallis test. Incidence of stroke, MI or death was estimated using Kaplan Meier analysis and compared by log rank test. eGFR ¼ Estimated glomerular filtration rate using the Lund-Malmö or Modification of Diet in Renal Disease isotope-dilution mass spectrometry aligned (MDRD) formulae; CHD ¼ Coronary heart disease; ACE ¼ angiotensin converting enzyme; ARB ¼ angiotensin receptor blocker. Data for the 12 patients that were reclassified to a less advanced eGFR category (5 to eGFR 90; 6 to eGFR 60e89; 1 to eGFR 45e59 ml/min/1.73 m2) by the Lund-Malmö formula in comparison to the MDRD equation are not shown.
295
296
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297
failure [36,37]. These previous studies, including a large metaanalysis of over one million adults, have reported that eGFR categorised by the CKD-EPI formula is more accurate at predicting adverse events than eGFR categorised by the MDRD formula [36,37]. Previous investigations that have assessed the association of eGFR with outcome in PVD patients however have almost always employed the MDRD formula [12e15,17e20]. We identified only one study that used the CKD-EPI formula to calculate eGFR as part of assessing the association of CKD with community outcomes in patients with PVD [11]. In that investigation eGFR was associated with outcome in 383 patients after undergoing endovascular aortic aneurysm repair. The current study suggests that research investigators and clinicians calculating eGFR as part of risk prediction in patients with PVD should consider using either the CKD-EPI or Lund-Malmö formulae. We identified no previous studies assessing the ability of eGFR categorised by the Lund-Malmö formula to predict major adverse events in any patient groups. This formula is not commonly used to assess eGFR however a recent study suggested that it gave the most accurate prediction of actual GFR in patients with late stage CKD [21]. In the current study eGFR categorised by the Lund-Malmö formula was equally as good at predicting major adverse events as eGFR categorised by the CKD-EPI formula. eGFR categorised by the Lund-Malmö formula was significantly better at predicting major adverse events than eGFR categorised by the MDRD formula. The Lund-Malmö formula defines substantially more patients to have CKD than both the CKD-EPI and MDRD formulae. Approximately an extra 15% of patients were categorised as having an eGFR <90 ml/ min/1.73 m2 and whether this is appropriate is controversial [10]. Overall our findings suggest both the CKD-EPI and Lund-Malmö formulae are appropriate to use in assessing risk of major adverse events in PVD patients. Other considerations apart from the accuracy of predicting major adverse events should also be considered before deciding on which formula to employ. These may include ability to compare with other studies, ease of calculating eGFR, availability of pre-calculated eGFR from local pathology departments and ability to predict other events such as progression to end stage renal failure. We did not examine these factors in the current study. A review of previous publications however suggests that the CKD-EPI formula is in more common use than the LundMalmö formula and many CKD foundations, such as Kidney Health Australia, recommend use of the CKD-EPI formula [38]. Taken together these factors favour the current use of the CKD-EPI formula in predicting eGFR in PVD patients. Our findings confirmed those from previous mainly smaller studies demonstrating an important increased risk of major adverse events in patients with PVD that have CKD reflected by reduced eGFR [12e20]. In the current study outcome was particularly poor for PVD patients with eGFR <45 with the estimated increased risk for patients categorised as eGFR 30e44, 15e29 and <15 ml/min/1.73 m2 approximately 1.5, 2 and 3-fold, respectively. The current study has a number of strengths and limitations. The sample size was relatively large compared to studies that have previously examined the association of CKD with outcome in PVD patients, although for some subgroups, such as those with eGFR <15 ml/min/1.73 m2 the numbers were small. The patients included in the current study presented with a range of different PVDs with varying severities of problems making the group representative of the overall population of patients with PVD but also quite heterogeneous. Thus our findings may have less applicability to very homogenous groups of patients with PVD such as those with lower limb athero-thrombosis alone. We examined three eGFR formulae although a number of others are available, such as the Cockroft-Gault, thus we can make no comment regarding the ability of other formulae to predict major adverse
events. We assessed eGFR only using serum creatinine and did not use other methods of assessing renal function such as serum cystatin C. Furthermore, proteinuria was not examined within the current study. Finally we did not assess the influence of amputation on eGFR. In conclusion the current study suggests that eGFR categorised by the CKD-EPI and Lund-Malmö formulae are equally effective at predicting major adverse events in PVD patients and their use is preferable to categorisation by the MDRD formula. Support and financial disclosure declaration Funding from the Queensland Government, the Townsville Hospital Private Practice Fund and National Health and Medical Research Council supported this work. JG holds a Practitioner Fellowships from the National Health and Medical Research Council, Australia (1019921). JG holds a Senior Clinical Research Fellowship from the Office of Health and Medical Research, Queensland. The authors have no conflict of interests relevant to this article. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atherosclerosis.2013.11.034. References [1] Eckardt KU, Coresh J, Devuyst O, et al. Evolving importance of kidney disease: from subspecialty to global health burden. Lancet 2013;382:158e69. [2] Gansevoort RT, Correa-Rotter R, Hemmelgarn BR, et al. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. Lancet 2013;382:339e52. [3] Layden J, Michaels J, Bermingham S, Higgins B, Guideline Development Group. Diagnosis and management of lower limb peripheral arterial disease: summary of NICE guidance. BMJ 2012;345:e4947. [4] Rooke TW, Hirsch AT, Misra S, et al. 2011 ACCF/AHA focused update of the guideline for the management of patients with peripheral artery disease (updating the 2005 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2011;58:2020e45. [5] European Stroke Organisation, Tendera M, Aboyans V, Bartelink ML, et al., ESC Committee for Practice Guidelines. ESC guidelines on the diagnosis and treatment of peripheral artery diseases: document covering atherosclerotic disease of extracranial carotid and vertebral, mesenteric, renal, upper and lower extremity arteries: the Task Force on the diagnosis and treatment of peripheral artery diseases of the European Society of Cardiology (ESC). Eur Heart J 2011;32:2851e906. [6] Nicolaides AN, Allegra C, Bergan J, et al. Management of chronic venous disorders of the lower limbs: guidelines according to scientific evidence. Int Angiol 2008;27:1e59. [7] Wattanakit K, Folsom AR, Selvin E, Coresh J, Hirsch AT, Weatherley BD. Kidney function and risk of peripheral arterial disease: results from the Atherosclerosis Risk in Communities (ARIC) Study. J Am Soc Nephrol 2007;18:629e36. [8] Garimella PS, Hart PD, O’Hare A, DeLoach S, Herzog CA, Hirsch AT. Peripheral artery disease and CKD: a focus on peripheral artery disease as a critical component of CKD care. Am J Kidney Dis 2012;60:641e54. [9] Florkowski CM, Chew-Harris JS. Methods of estimating GFRedifferent equations including CKD-EPI. Clin Biochem Rev 2011;32:75e9. [10] Moynihan R, Glassock R, Doust J. Chronic kidney disease controversy: how expanding definitions are unnecessarily labelling many people as diseased. BMJ 2013;347:f4298. [11] Saratzis A, Sarafidis P, Melas N, Saratzis N, Kitas G. Impaired renal function is associated with mortality and morbidity after endovascular abdominal aortic aneurysm repair. J Vasc Surg 2013;58:879e85. [12] Lacroix P, Aboyans V, Desormais I, et al., COPART Investigators. Chronic kidney disease and the short-term risk of mortality and amputation in patients hospitalized for peripheral artery disease. J Vasc Surg 2013;58:966e71. [13] Liang KW, Kuo HN, Lee WL, et al. Different mid-term prognostic predictors of major adverse events in diabetic and nondiabetic peripheral artery disease presenting with critical limb ischemia. Angiology 2013 Feb 6 [Epub ahead of print]. [14] Romero JM, Bover J, Fite J, Bellmunt S, Dilmé JF, Camacho M, et al. The modification of diet in renal disease 4-calculated glomerular filtration rate is a better prognostic factor of cardiovascular events than classical cardiovascular risk factors in patients with peripheral arterial disease. J Vasc Surg 2012;56: 1324e30.
J. Golledge et al. / Atherosclerosis 232 (2014) 289e297 [15] Kumakura H, Kanai H, Aizaki M, et al. The influence of the obesity paradox and chronic kidney disease on long-term survival in a Japanese cohort with peripheral arterial disease. Vasc Med 2011;16:422e7. [16] Ohrlander T, Dencker M, Dias NV, Gottsäter A, Acosta S. Cardiovascular predictors for long-term mortality after EVAR for AAA. J Vasc Surg 2010;52: 110e7. [17] Jaffery Z, Greenbaum AB, Siddiqui MF, et al. Predictors of mortality in patients with lower extremity peripheral arterial disease: 5-year follow-up. J Interv Cardiol 2009;22:564e70. [18] Pasqualini L, Schillaci G, Pirro M, et al. Renal dysfunction predicts long-term mortality in patients with lower extremity arterial disease. J Intern Med 2007;262:668e77. [19] Owens CD, Ho KJ, Kim S, et al. Refinement of survival prediction in patients undergoing lower extremity bypass surgery: stratification by chronic kidney disease classification. J Vasc Surg 2007;45:944e52. [20] O’Hare AM, Bertenthal D, Shlipak MG, Sen S, Chren MM. Impact of renal insufficiency on mortality in advanced lower extremity peripheral arterial disease. J Am Soc Nephrol 2005;16:514e9. [21] Evans M, van Stralen KJ, Schön S, et al., On the behalf of the ERA-EDTA Registry and the Swedish Renal Registry collaboration. Glomerular filtration rateestimating equations for patients with advanced chronic kidney disease. Nephrol Dial Transplant 2013;28:2518e26. [22] Golledge J, Cronin O, Iyer V, Bradshaw B, Moxon JV, Cunningham MA. Body mass index is inversely associated with mortality in patients with peripheral vascular disease. Atherosclerosis 2013;229:549e55. [23] Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373e9. [24] Palamuthusingam D, Quigley F, Golledge J. Implications of the finding of no significant carotid stenosis based on data from a regional Australian vascular unit. Ann Vasc Surg 2011;25:1050e6. [25] Parr A, Buttner P, Shahzad A, Golledge J. Relation of infra-renal abdominal aortic calcific deposits and cardiovascular events in patients with peripheral artery disease. Am J Cardiol 2010;105:895e9. [26] Golledge J, Jayalath R, Oliver L, Parr A, Schurgers L, Clancy P. Relationship between CT anthropometric measurements, adipokines and abdominal aortic calcification. Atherosclerosis 2008;197:428e34. [27] Parr A, McCann M, Bradshaw B, Shahzad A, Buttner P, Golledge J. Thrombus volume is associated with cardiovascular events and aneurysm growth in patients who have abdominal aortic aneurysms. J Vasc Surg 2011;53:28e35.
297
[28] Magee R, Quigley F, McCann M, Buttner P, Golledge J. Growth and risk factors for expansion of dilated popliteal arteries. Eur J Vasc Endovasc Surg 2010;39: 606e11. [29] Porter JM, Moneta GL. Reporting standards in venous disease: an update. International consensus committee on chronic venous disease. J Vasc Surg 1995;21:635e45. [30] Eknoyan G, Hostetter T, Bakris GL, et al. Proteinuria and other markers of chronic kidney disease: a position statement of the national kidney foundation (NKF) and the national institute of diabetes and digestive and kidney diseases (NIDDK). Am J Kidney Dis 2003;42:617e22. [31] Van Campenhout A, Moran CS, Parr A, et al. Role of homocysteine in aortic calcification and osteogenic cell differentiation. Atherosclerosis 2009;202: 557e66. [32] Björk J, Grubb A, Sterner G, Nyman U. Revised equations for estimating glomerular filtration rate based on the Lund-Malmö Study cohort. Scand J Clin Lab Invest 2011;71:232e9. [33] Anderson JL, Adams CD, Antman EM, et al., 2011 Writing Group Members; ACCF/AHA Task Force Members. 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 guidelines for the management of patients with unstable Angina/Non-ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011;123:e426e579. [34] Wright L, Hill KM, Bernhardt J, et al., National Stroke Foundation Stroke Guidelines Expert Working Group. Stroke management: updated recommendations for treatment along the care continuum. Intern Med J 2012;42: 562e9. [35] Pencina MJ, D’Agostino Sr RB, D’Agostino Jr RB, Vasan RS. Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond. Stat Med 2008;27:157e72. [36] Matsushita K, Mahmoodi BK, Woodward M, et al., Chronic Kidney Disease Prognosis Consortium. Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. J Am Med Assoc 2012;307:1941e51. [37] Oh J, Kang SM, Hong N, et al., KorHF Registry. The CKD-EPI is more accurate in clinical outcome prediction than MDRD equation in acute heart failure: data from the Korean Heart Failure (KorHF) registry. Int J Cardiol 2013;167:1084e7. [38] Kidney Health Australia. GFR calculator. Available from: http://www.kidney. org.au/HealthProfessionals/GFRCalculatorCKDEPI/tabid/803/Default.aspx. [accessed 30.08.13].