Microalbuminuria and chronic kidney disease as risk factors for cardiovascular disease

The earlier observations on the association between renal failure and increased rates for cardiovascular (CV) complications and death date back several decades 1 , 2 . Similarly, it has long been noted that subclinical elevations of urinary albumin excretion (UAE) are related to higher risk of subsequent development of clinical nephropathy in patients with diabetes mellitus 3 , 4 , and that increased UAE was associated with higher risk for CV events and mortality in both diabetic and non-diabetic individuals 4–7 . Since then, a substantial amount of data have accumulated, providing solid evidence that both the decline in renal function and the elevation in UAE are independently associated with increased risk for cardiovascular disease (CVD). This article reviews the roles of microalbuminuria (MA) and chronic kidney disease (CKD) as risk factors for CVD, discussing the progress of the epidemiological and clinical evidence on the field. In the first description of MA as a predictor of nephropathy in patients with type 1 diabetes in 1982, Viberti et al . 3 defined MA as the overnight UAE rate between 30 and 140 mg/min and clinical nephropathy as an ‘Albustix’ stick positive for proteinuria. Since then, the definition of MA has been refined to include all the possible methods of UAE measurement. Currently, MA is defined as UAE between 30 and 300 mg/day, if measured in a 24 h urine collection, 20–200 μg/min, if measured in a timed urine collection or 30–300 mg/g, if measured with the use of urinary albumin to creatinine ratio (UACR) in a spot urine collection. Any urinary albumin value below these limits is considered as normal UAE, whereas any value above them reflects the presence of macroalbuminuria or clinical proteinuria 8 , 9 . CKD is defined either as kidney damage, as confirmed by kidney biopsy or markers of damage, or as the presence of glomerular filtration rate (GFR) under the level of 60 ml/min/1.73 m 2 , each for a period greater than 3 months 10 . The clinical syndrome of CKD is divided in terms of severity in five stages, which are based on the level of GFR, irrespective of the cause of kidney damage. Decreased kidney function starts at a GFR below 89 ml/min/1.73 m 2 and is considered to be pronounced if the GFR is 55 years of age 11 , 17 , 18 . Without specific intervention, the rate of progression to diabetic nephropathy (i.e. development of macroalbuminuria) in patients with type 2 diabetes and MA would be 5% per year, whereas in patients with type 1 diabetes and MA, it would be 7.5% per year 3 , 4 . Subsequent end-stage renal disease (ESRD) occurs at a rate of 1% annually in type 2 diabetes patients, while the risk for those with type 1 diabetes approaches 75% after 20 years 8 , 19 , 20 . It should be noted, however, that the rate of nephropathy progression, as well as CV risk, is far lower in diabetic patients who have tight control of glucose and BP levels early in the course of their disease 21 . Among non-diabetic individuals with essential hypertension, the prevalence of MA varies widely from 5 to 40% 22–24 . The reason for this high variability lies again in differences among the population studies in the actual BP levels, and other factors known to influence MA levels ( Table 1 ). Factors known to influence the development of microalbuminuria in subjects without diabetes Factors known to influence the development of microalbuminuria in subjects without diabetes During the past decades, CKD has grown to represent a world-wide public health problem. Data from National Health and Nutrition Examination Survey (NHANES) III (1988–94) suggest that the prevalence of CKD in the adult population of the US was roughly 11%, thus about 20 million adult individuals were estimated to suffer from CKD, among which more than 8 million have CKD of at least stage 3 and 300 000 require dialysis treatment 10 , 25 . The incidence and prevalence of ESRD is continuously increasing. In 2002, the number of patients with ESRD rose up to more than 430 000 and the incident rate of ESRD increased to 333 new cases per million people, a number almost four times higher than that in 1980 26 . The prevalence of ESRD was even higher in Japan, and generally lower in Europe, although wide differences were observed among European countries 27 . In spite of those differences, however, data from the ERA–EDTA registry also clearly suggest an upward trend for ESRD incidence in all the countries examined 27 , 28 . Diabetic nephropathy remains by far the number one cause for ESRD in western countries with a prevalence rising to around 40% of ESRD patients 8,10,26,29 . Twenty-four hour or other timed collections were the traditional way to measure UAE but measuring UACR in a spot collection of morning urine in the fasting state is currently recommended as a simple, quick and comparatively accurate way of determining albuminuria 8,10,30,31 . The use of this approach, however, requires knowledge of the factors that can affect spot UACR measurement ( Table 2 ). It is important to note that the range of UAE is about 25% lower during sleep than while awake and MA can exhibit a daily intra-individual variation between 40 and 100% 32–34 . This can be largely attributed to biological variations due to inflammation associated with small injury, toothaches, etc. as well as changes in dietary sodium and protein intake 35 . Caution is required when interpreting the UACR in patients with higher muscle mass, i.e. males or African-Americans, as these populations have higher levels of creatinine excretion 36 . That is why different levels for MA in males and females have been proposed 9 . The imprecise nature of MA and creatinine measurements require at least three measures to be made over a period of 2–3 months before determining the actual UACR for a particular individual 30 . Ideally, these measurements should be obtained in the fasting state and collected from the first morning void, to avoid the effect of any physical activity during the day 37 . Factors that can affect the measurement of urine albumin to creatinine rate in a spot specimen Factors that can affect the measurement of urine albumin to creatinine rate in a spot specimen Serum creatinine has been long considered an imprecise index of renal function, since it is affected by various factors, apart from creatinine filtration; patients with mild renal impairment can thus have normal or near normal levels 38 , 39 . This is why stages of CKD are defined with the use of GFR levels, which is considered the best index of renal function in both health and disease 10 . Most accurate estimations of GFR require the measurement of renal clearance of inulin, which is cumbersome and time-consuming, so is the use of various radioactive substances, which also gives accurate results 40 , but seems difficult for everyday clinical practice. To overcome these limitations, a number of equations using serum creatinine and demographic variables have been developed to estimate GFR through creatinine clearance, such as the Cockcroft–Gault formula, the Modification of Diet in Renal Disease (MDRD) study formula and others 41 , 42 . The MDRD study formula seems to provide a much more accurate estimate of GFR than other commonly used equations 42 and therefore it is currently recommended for GFR determination 10 , 31 . Yudkin et al . 43 were the first to report cross-sectional associations between MA and the prevalence of coronary artery disease (CHD) and peripheral vascular disease (PVD). Since then, several studies, either in the general population or in high-risk individuals, have reported the associations of MA with established CV risk factors (age, hypertension, hyperglycaemia, obesity, high-total and LDL-cholesterol, high triglycerides, low HDL-cholesterol, smoking) as well as emerging CV risk factors and other unfavourable conditions (high C-reactive protein, insulin resistance, hyperinsulinaemia, endothelial dysfunction, hyperhomocysteinaemia, high-fibrinogen levels and others) 14 , 16 , 17 , 44–60 . Furthermore, MA has been long suggested to be an independent risk factor for CV morbidity and mortality in various populations, either with or without diabetes mellitus. In patients with type 1 or type 2 diabetes, early observations in small cohorts of subjects followed for periods up to many years linked the presence of MA with increased risk of CV events and overall or CV death, in parallel to an enormous elevation of the risk for future development of clinical proteinuria 4 , 6 , 7 . Subsequent prospective cohort studies further supported these findings, showing independent associations between MA and CV or total mortality 11 , 61 , 62 . Similarly, in non-diabetic individuals, MA was also found to be a strong predictor of future CV morbidity and mortality, independent of major CV risk factors of 3.5 63 , 64 . However, the best evidence of the association between MA and the risk for CVD comes from post-hoc analyses of long-term clinical trials, as well as from recent large prospective cohort studies. In individuals with a history of CVD or diabetes of the Heart Outcomes Prevention Evaluation (HOPE) trial, MA was associated with an adjusted relative risk of 1.83 for major CV events, 2.09 for all-cause mortality and 3.23 for hospitalization for congestive heart failure 65 , with similar relative risks in subjects with or without diabetes. Notably, the risk for the primary endpoint started from UACR levels well below the cut-off for MA and increased continuously. For every 0.4 mg/mmol increase in the UACR level, the adjusted hazard of major CV events increased by 5.9%. Similarly, in a post-hoc analysis of the Losartan Intervention For Endpoint reduction in hypertension (LIFE) study 66 , the risk for the primary composite endpoint (CV death, fatal and non-fatal stroke, and fatal and non-fatal myocardial infarction) increased continuously from very low levels of UAE. In non-diabetic patients, for every 10-fold increase in UACR, hazard ratios for the composite endpoint increased by 57%, for CV mortality by 97.7%, for all-cause mortality by 75.2%, for stroke by 51.0% and for myocardial infarction by 45%. The results for diabetic patients were similar, with the exception of myocardial infarction, which was not significant. In addition to the above, in a subpopulation of the NHANES II study, examined between 1976 and 1980 and followed for 16 years, the adjusted relative hazards for CV and all-cause mortality were 1.57 and 1.64 for subjects with urinary protein levels of 30–299 mg/dl and 1.77 and 2.00, respectively, in subjects with urinary protein levels ≥300 mg/dl, compared with the individuals with levels of <30 mg/dl 67 . A series of studies on subpopulations of 20 000–23 600 individuals from the population-based European Prospective Investigation into Cancer and Nutrition study, Norfolk, UK (the EPIC-Norfolk Study) further support the above. In these studies, after mean follow-up of 6.2–7.2 years, the presence of MA at baseline was independently associated with a significant greater risk of 36% for incident CHD 68 , 49% for stroke 69 , 103% for CV mortality and 48% for all-cause mortality 70 . Notably, the risk for CV events associated with the presence of macroalbuminuria was again even higher 68 , 69 . In a very recent prospective study, MA at baseline was also independently associated with future CV events after an average follow-up of 42.5 months 71 . In contrast to the above, it is still uncertain whether the reduction of MA is linked to a reduction in CV endpoints. In the LIFE trial, the group of subjects with the lowest CV event rate also had the greatest reduction in UAE from baseline and one-fifth of the difference in favour of Losartan on the primary composite endpoint was attributed to the greater reduction in albuminuria on Losartan 72 . In the Prevention of Renal and Vascular End-stage Disease Intervention Trial (PREVEND IT), subjects with MA were randomized to fosinopril 20 mg/day or matching placebo and pravastatin 40 mg/day or matching placebo. During a mean follow-up of 46 months, fosinopril reduced UAE by 26% but the reduction in CV mortality or CV hospitalization in patients treated with this agent showed only a trend towards statistical significance 73 . This study, however, was limited by the low overall number of CV events, as well as submaximal doses of fosinopril. It has been hypothesized that MA can also serve as a marker for subclinical CV organ damage. In various previous studies, MA has been associated with larger left ventricular mass and higher degrees of left-ventricular hypertrophy 14 , 74–77 . Even among hypertensive patients who are normoalbuminuric, those with higher levels of UAE have greater left ventricular and more left-ventricular hypertrophy . In the LIFE study, UAE was with left-ventricular hypertrophy at baseline and after 1 year of treatment independently of glucose and age, and the changes in these during the 1 year period were also independent of BP and glucose changes . MA has also been associated with the of with the use of of the in the artery 14 , , , , or the prevalence of vascular changes 14 , . as patients with diabetes, hypertension or other risk factors for CVD progress to major CV events through an that is by the presence of subclinical organ peripheral and this not only but also the of major events , for MA can be an for high-risk of UAE is also considered a very for of the severity of any The level of UAE is to the severity of many such as and . The severity of conditions and is also associated with the level of UAE. For MA was found to be an early in patients with myocardial infarction and to the severity of the , whereas in patients with and the level of UACR after was related to the severity of muscle during . the earlier observations that patients with ESRD had rates for CV complications and death, were made more than 30 years 1 , 2 as Among patients in the in the of those with CKD CVD over a period compared with only of those without CKD, and the prevalence of heart failure was four times higher in CKD patients compared with those without CKD 26 . In ESRD patients treated by the risk of CVD was with CV mortality rates in various studies times higher than and in the general population . to CVD is by far the cause of death in ESRD patients on both of the 26 , . In several of the above studies MA with CVD risk, was a association between the level of UAE and the risk for CVD macroalbuminuria or clinical proteinuria was associated with a higher risk for CV morbidity and mortality than MA 61 , . Since the presence of macroalbuminuria is a of nephropathy and is associated with of kidney function 9 , 10 , 89 , the for the increase in CV risk in patients be different from those in patients with the CV risk in patients with macroalbuminuria be attributed not only to the presence of a vascular injury, as in the of MA, but also to the of a number of factors which can also In parallel to the presence of many of the traditional diabetes, obesity, or (i.e. CV risk factors in patients with CKD 9 , as GFR below 60 ml/min/1.73 m 2 and of the kidney to a number of factors related to renal function and which can further to CVD also , , providing for the independent between renal function and CV These factors when the GFR below ml/min/1.73 m 2 and at a GFR of 30 ml/min/1.73 m 2 . The important of these factors to be in and related and from reduced of in and have been associated with increased vascular and CV risk, while recent evidence that reduced levels of also to increased CV mortality in CKD patients . the other through various on the CV left-ventricular is also associated with CV in CKD patients . evidence that both and of levels are associated with reduction of CV events , . Since the very first observations in ESRD patients 1 , 2 , studies have the association between renal function and CV or overall mortality in populations with or without Most of these studies suggested that mild to elevations in serum creatinine levels were associated with increased risk of CV events or death , as well as of death from any cause 5 , , , , whereas a not confirm these results . A major of these studies was the use of only serum creatinine levels to the level of renal function 5 , , with different cut-off to the presence or the stages of CKD 5 , , , as well as the use of of estimated kidney function 5 , , , , . were also by small of individuals with CKD 5 , , , , and others were to the study of populations patients with CVD, , . it is only very that studies either on the general population 67 , , or in patients with CVD , the association between the of kidney function and CV risk using the level of GFR and more In the study in the of the NHANES II population 67 , mortality rates were and of follow-up among with estimated GFR of and In individuals with estimated GFR of had a higher risk of for death from CVD and for overall mortality compared with subjects GFR In the population-based In study , subjects with GFR of ml/min/1.73 m 2 had a adjusted higher risk for CV events and subjects with GFR of ml/min/1.73 m 2 had a higher risk compared with subjects with GFR of ml/min/1.73 m 2 , after years of each 10 ml/min/1.73 m 2 lower GFR was associated with a significant adjusted hazard ratio of and for CVD, CVD and CVD, Similarly, in a cohort of almost subjects over 65 years of age, followed for 5 years, each 10 ml/min/1.73 m 2 lower GFR was independently associated with a hazard ratio of for CVD, for CVD, for CVD and for all-cause mortality . A recent prospective cohort study on more than million individuals in the US seems to provide the best evidence on the . a follow-up of years, which to over 3 million of the adjusted hazard ratio for CV events was for individuals with a GFR ml/min/1.73 m 2 , for those with GFR ml/min/1.73 m 2 , for those with GFR ml/min/1.73 m 2 and for individuals with GFR ml/min/1.73 m 2 compared with the subjects with GFR 60 ml/min/1.73 m 2 , who were used as the The adjusted risk of overall mortality and hospitalization followed the Notably, the for the adjusted risks were very due to the enormous thus the results provide very accurate of the actual studies these on independent associations between the GFR estimated with the MDRD formula and CV risk also in patients with CVD , . In 14 patients with myocardial infarction by heart left-ventricular dysfunction, or from the population of in Trial the adjusted hazard ratio for the composite endpoint from CVD, congestive heart myocardial infarction, after and was for individuals with GFR ml/min/1.73 m 2 , for those with GFR ml/min/1.73 m 2 and for those with GFR ml/min/1.73 m 2 , respectively, compared with subjects of the group with GFR ml/min/1.73 m 2 . The adjusted risk from overall death was also with GFR, while for GFR below a higher renal function level ml/min/1.73 m 2 each was associated with a increase in the risk of death and non-fatal CV Similarly, a recent analysis examined the association of kidney function with CVD in patients with chronic heart from the population of the in Heart of in and . a follow-up of months, the adjusted hazard ratio for the primary of CV death or hospitalization for congestive heart failure was for patients with GFR ml/min/1.73 m 2 and for those with GFR ml/min/1.73 m 2 compared with patients with GFR 60 ml/min/1.73 m 2 and similar was the association between GFR and all-cause Notably, this effect of GFR was not related to the level of left-ventricular which was also found to be an independent predictor of that examined the of renal after coronary artery in patients with coronary artery studies on the have that both ESRD , and renal not renal are associated with increased morbidity and mortality after However, the latter studies also patients with renal by the examined populations into on the of serum creatinine level and not measuring renal function as a . recent studies that used GFR as a measure of renal function, estimated either with the Cockcroft–Gault , or the MDRD formula , , have clearly that reduced GFR was independently associated with morbidity and mortality as well as as in the general in the latter studies the between renal and mortality risk was not but a increase when GFR below a of about 60 ml/min/1.73 m 2 , . evidence a of the between MA or CKD and the risk for MA and CKD are independently associated with a higher risk for CV events, as well as CV and overall mortality in the general population and in patients with other CV risk factors or the and clinical significance of these associations to be these suggest that determination of UAE and renal function levels should be in clinical for overall risk at least in individuals with high CVD The for of those variables is further supported by the that determination of MA with the use of UACR and level of renal function with of GFR are to in the clinical and the results are should in the populations that would from these measurements and future should provide for their clinical of .