Journal of Human Hypertension (2004) 18, 453–459 & 2004 Nature Publishing Group All rights reserved 0950-9240/04 $30.00 www.nature.com/jhh
ORIGINAL ARTICLE
Albuminuria predicts cardiovascular events independently of left ventricular mass in hypertension: a LIFE substudy MH Olsen1, K Wachtell1, JN Bella2, V Palmieri2, E Gerdts3, G Smith4, MS Nieminen5, B Dahlo¨f6, H Ibsen1 and RB Devereux2 1
Glostrup University Hospital, Glostrup, Denmark; 2The Weill Medical College of Cornell University, New York, NY, USA; 3Haukeland Hospital, Bergen, Norway; 4Helsinki University Hospital, Helsinki, Finland; 5 Ullevaal University Hospital, Oslo, Norway; 6Sahlgrenska University Hospital-O¨stra, Sweden
We wanted to investigate whether urine albumin/creatinine ratio (UACR) and left ventricular (LV) mass, both being associated with diabetes and increased blood pressure, predicted cardiovascular events in patients with hypertension independently. After 2 weeks of placebo treatment, clinical, laboratory and echocardiographic variables were assessed in 960 hypertensive patients from the LIFE Echo substudy with electrocardiographic LV hypertrophy. Morning urine albumin and creatinine were measured to calculate UACR. The patients were followed for 6074 months and the composite end point (CEP) of cardiovascular (CV) death, nonfatal stroke or nonfatal myocardial infarction was recorded. The incidence of CEP increased with increasing LV mass (below the lower quartile of 194 g to above the upper quartile of 263 g) in patients with UACR below (6.7, 5.0, 9.1%) and above the median value of 1.406 mg/ mmol (9.7, 17.0, 19.0%***). Also the incidence of CV death increased with LV mass in patients with UACR
below (0, 1.4, 1.3%) and above 1.406 mg/mmol (2.2, 6.4, 8.0%**). The incidence of CEP was predicted by logUACR (hazard ratio (HR) ¼ 1.44** for every 10-fold increase in UACR) after adjustment for Framingham risk score (HR ¼ 1.05***), history of peripheral vascular disease (HR ¼ 2.3*) and cerebrovascular disease (HR ¼ 2.1*). LV mass did not enter the model. LogUACR predicted CV death (HR ¼ 2.4**) independently of LV mass (HR ¼ 1.01* per gram) after adjustment for Framingham risk score (HR ¼ 1.05*), history of diabetes mellitus (HR ¼ 2.4*) and cerebrovascular disease (HR ¼ 3.2*). *Po0.05, **Po0.01, ***Po0.001. In conclusion, UACR predicted CEP and CV death independently of LV mass. CV death was predicted by UACR and LV mass in an additive manner after adjustment for Framingham risk score and history of CV disease. Journal of Human Hypertension (2004) 18, 453–459. doi:10.1038/sj.jhh.1001711 Published online 15 April 2004
Keywords: Albuminuria; left ventricular hypertrophy; hypertension; cardiovascular events; prognosis
Introduction High urine albumin–creatinine ratio (UACR) as well as left ventricular (LV) mass has been associated with increased cardiovascular (CV) risk in population studies.1,2 We have recently extended these findings to patients with hypertension and electrocardiographic LV hypertrophy.3 Albuminuria is associated with atherosclerotic risk factors like ageing, diabetes and high blood pressure4,5 and is thought to reflect microvascular damage not only in the glomeruli but also in the systemic vasculature6 with increasing risk of lipid insudation and development of atherosclerosis.7 Increased LV mass is
associated with the same CV risk factors and is associated with vascular hypertrophy.8 Although they are related to the same CV risk factors, increased UACR may reflect subclinical disease in the small arteries and capillaries, whereas LV mass reflects subclinical disease in the heart and probably in the large arteries. Therefore, we wanted to investigate whether high UACR supplemented high LV mass in predicting the composite end point (CEP) of CV death, nonfatal stroke or nonfatal myocardial infarction in patients with hypertension LV hypertrophy.
Methods Correspondence: Dr. MH Olsen, Department of Clinical Physiology and Nuclear Medicine, Glostrup University Hospital, DK2600 Glostrup, Denmark. E-mail:
[email protected] Published online 15 April 2004
Study design
As part of the LIFE trial, an echocardiographic substudy enrolled 960 participants, approximately
Albuminuria, LV mass and cardiovascular events MH Olsen et al 454
11% of the parent trial population. As described elsewhere,9 patients aged 55–80 years with mean seated blood pressures in the range 160–200 mmHg systolic and/or 95–115 mmHg diastolic after 1 and 2 weeks on placebo treatment. Patients had electrocardiographic LV hypertrophy by either the Cornell voltage–duration product (RaVL þ SV3 þ 6 mV in women 42440 ms mV) or Sokolow–Lyon voltage criteria (SV1 þ RV5/6438 mV). Patients were recruited into the LIFE echocardiography substudy at sites in Denmark, Finland, Iceland, Norway, Sweden, the United Kingdom and the United States. CV disease was defined as coronary artery disease (history of either angina or myocardial infarction), peripheral vascular disease (claudication or peripheral arterial angioplasty) or cerebrovascular disease (stroke or transient ischaemic attack). The patients were randomized to atenolol- or losartan-based antihypertensive regimens for 4450 patient-years of followup. The predefined end point of the LIFE study, a composite CV end point consisting of CV death, fatal and nonfatal myocardial infarction, and fatal and nonfatal stroke, was registered and adjudicated by an independent end point committee. Echocardiographic methods
Studies were performed using echocardiographs with M-mode, two-dimensional and pulsed, continuous and colour-flow Doppler capabilities, as previously described.10,11 Studies were sent to the Cornell Medical Center Echocardiography Reading Center for blinded interpretation by experienced technician and physician readers. LV internal dimension and wall thickness were measured at end-diastole and end-systole following American Society of Echocardiography (ASE) recommendations12 on up to three cardiac cycles. When optimal orientation of the M-mode cursor could not be obtained, correctly oriented linear dimension measurements were made using two-dimensional imaging by the leading-edge ASE convention.13 Calculation of derived variables
End-diastolic LV dimensions were used to calculate LV mass by a formula shown to yield values closely related to necropsy LV weight (r ¼ 0.90, Po0.001) and good reproducibility (RHO ¼ 0.93) between echocardiograms in hypertensive patients with LV hypertrophy.14 LV mass was considered as an unadjusted variable and also after traditional normalization for body surface area.10,11 Standard methods were used to calculate endocardial fractional shortening and ejection fraction.15 Stroke volume was calculated by the invasively validated aortic annular-leading edge method as (aortic annular cross-sectional area) (Doppler time velocity integral of aortic annular flow)16 and used to calculated cardiac output and total peripheral Journal of Human Hypertension
resistance. The pulse pressure/stroke volume ratio was used as an indirect measure of systemic arterial stiffness.17 The primary approach used to assess myocardial contractile performance was examination of LV midwall fractional shortening (MWS) in relation to circumferential end-systolic stress (ESS) measured at LV minor axis.18 The formulae used to calculate MWS and midwall circumferential ESS have been previously described.15,18 To evaluate LV performance taking ESS into account, observed shortening was expressed as a percentage of the value predicted from circumferential ESS using equations derived from normal subjects.18 For convenience, this variable is termed stress-corrected MWS.19 A partition value of o89.2% for stress-corrected MWS was defined as subnormal, based on the 5th percentile in a previously described reference population of 280 normotensive, nondiabetic, normal weight men and women aged 18–88 years.20 Urine albumin excretion
Urine albumin concentration was determined by standard methods21,22 by using a turbidimetric method (Hitachi 717 Analyzer, Boehringer Mannheim)23,24 on a single urine specimen taken on the morning of electrocardiography. Both serum and urine creatinine were analysed using the Jaffe´ reaction without deproteinizing and then quantified by a photometric method (Hitachi 717 Analyzer, Boehringer Mannheim). The urine-albumin concentration, expressed as a ratio to the urinary creatinine concentration, was used as a measure of glomerular vascular permeability to allow for urine dilution. We defined microalbuminuria as UACR X3.5 mg/mmol and macroalbuminuria as UACR X 35.0 mg/mmol.25 Statistical analyses
Data management and analysis were performed using the SPSS 10.1 (SPSS, Chicago, IL, USA) software. Data are presented as mean7s.d. for continuous variables and proportions for categorical variables. Unpaired Student’s t-test was used to determine differences in continuous variables between groups. Pearson’s w2-test was used to compare differences in categorical variables between groups. When performing stepwise, backward multiple linear regression analyses calculating the standardized regression quotient for each parameter (r) all basic characteristics (ie age, gender, weight, height, systolic pressure, LV mass, glucose and cholesterol levels) entered the model. Hazard ratios (HRs) and their 95% confidence intervals (in brackets) for CV events were assessed using Cox regression analysis adjusting for potential confounders (ie age, gender, body mass index, systolic pressure, glucose and cholesterol levels, prior CV disease and treatment allocation). Nonsignificant factors were excluded
Albuminuria, LV mass and cardiovascular events MH Olsen et al 455
stepwise. Two-tailed Po0.05 indicated statistical significance.
the LV mass groups (24.6% (58/236), 29.0% (136/ 469) and 37.6% (88/234), Po0.01) and in patients with micro- or macroalbuminuria (39.3% (77/196) and 53.7% (22/41) vs 25.9% (155/599), Po0.001).
Results At baseline, LV mass was obtained in 939 of the 960 patients and the patients were divided into three groups according to LV mass below, between or above the lowest (194 g) and highest (263 g) quartiles (Table 1). UACR was measured in 836 patients and microalbuminuria was found in 192 (23.4%) patients and macroalbuminuria in 40 (4.9%) patients. Patients with UACR above the median value of 1.406 mg/mmol had more CV risk factors (Table 2). The prevalence of micro- (19.3% (41/212), 19.5% (80/410) and 35.0% (75/214)) and macroalbuminuria (1.4% (3/212), 4.9% (20/410) and 8.4% (18/214)) increased throughout the three LV mass groups (Po0.001). High LV mass was in multiple regression analyses associated with higher weight (r ¼ 0.34, Po0.001) and height (r ¼ 0.34, Po0.001), male gender (r ¼ 0.22, Po0.001), older age (r ¼ 0.13, Po0.001), higher systolic blood pressure (r ¼ 0.12, Po0.001), and lower total cholesterol (r ¼ 0.07, Po0.001). LogUACR was associated with plasma glucose (r ¼ 0.20, Po0.001), systolic blood pressure (r ¼ 0.20, Po0.001), current smoking (r ¼ 0.14, Po0.001), LV mass index (r ¼ 0.12, P ¼ 0.001), age (r ¼ 0.08, Po0.05) and serum high-density lipoproteins (r ¼ 0.07, Po0.05). The prevalence of patients with known high CV risk due to history of coronary heart disease, peripheral vascular disease, cerebrovascular disease or diabetes increased throughout
CV end points
With increasing LV mass, the incidence of CEP (7.6% (18/236), 10.4% (49/469) and 16.2% (38/234), P ¼ 0.01) as well as CV death (0.8% (2/236), 3.6% Table 2 Characteristics of patients grouped by UACR UACRo1.406 mg/mmol Age (years) Height (cm) Weight (kg) BMI (kg/m2) Glucose (mmol/l) Cholesterol (mmol/l) HDL (mmol/l) Systolic BP (mmHg) Diastolic BP (mmHg) LV PWTD (mm) LVIDD (mm) LV RWT (ratio) LV mass (g) LVMI (mmHg) FS (%) MWS (%) IVRT (ms) Dec. time (ms) E/A-ratio (ratio)
UACRX1.406 mg/mmol
6576 16979 78713 27.274.0 5.671.3 6.171.1 1.5770.43 171714 9878 10.671.2 52.475.5 0.4170.07 226752 120723 33.575.6 15.672.1 116724 220766 0.8470.31
6777 17079 79715 27.474.7 6.573.0 5.971.1 1.4670.46 176713 99710 10.971.3 53.276.1 0.4270.07 243762 128728 33.575.2 15.372.1 115722 215764 0.8870.44
o0.001
o0.001 o0.01 o0.001 o0.001 o0.001 o0.05 o0.001 o0.001
Data are presented as mean values and standard deviations. Abbreviations are same as in Table 1.
Table 1 Characteristics of patients grouped by LV mass LV mass Age (years) Height (cm) Weight (kg) BMI (kg/m2) Glucose (mmol/l) Cholesterol (mmol/l) HDL (mmol/l) Systolic BP (mmHg) Diastolic BP (mmHg) LV PWTD (mm) LVIDD (mm) LV RWT (ratio) LV mass (g) LVMI (mmHg) FS (%) MWS (%) IVRT (ms) Dec. time (ms) E/A-ratio (ratio)
LV mass o194 g
194gpLV mass o263 g
LV massX263 g
6577 16478 71711 26.373.8 5.872.1 6.271.2 1.7070.52 172714 9779 9.870.8 57.973.1 0.4170.05 173718 99712 36.274.6 16.471.8 111720 217771 0.8470.28
6677 17079&&& 78712&&& 27.174.3& 6.172.4 6.071.1& 1.5070.42&&& 174714 9879 10.670.9&&& 52.674.1&&& 0.4170.06 227720&&& 121712&&& 33.575.2&&& 15.571.9&&& 114725&&& 216763 0.8770.38
6677 17578yyy 87715yyy 28.474.7yyy 6.172.5 5.771.1yyy 1.3770.38yyy 175713 99710 11.771.4yyy 58.575.6yyy 0.4170.08 310745yyy 154725yyy 30.176.2yyy 14.272.2yyy 122721yyy 216764 0.8970.44
ANOVA
*** *** *** *** *** * ** *** *** *** *** *** *** ***
Data are presented as mean values and standard deviations. Abbreviations: UACR, urine albumin/creatinine ratio; BMI, body mass index; HDL, high-density lipoproteins; BP, blood pressure; LV, left ventricular; PWTD, posterior wall thickness in diastole; IDD, internal end-diastolic diameter; RWT, relative wall thickness; FS, endocardial fractional shortening; MWS, midwall fractional shortening; IVRT, isovolumetric relaxation time; Dec. time, deceleration time; E, early left ventricular filling; A, atrial left ventricular filling. *Po0.05, **Po0.01, ***Po0.001. Journal of Human Hypertension
Albuminuria, LV mass and cardiovascular events MH Olsen et al 456
(17/469) and 6.0% (14/234), P ¼ 0.01) increased. The HRs increased also with increasing LV mass (CEP: HR ¼ 1.40 (0.81–2.40), NS and HR ¼ 2.27 (1.30– 2.98), Po0.01 and CV death: HR ¼ 4.24 (0.98–18.4), P ¼ 0.053 and HR ¼ 7.39 (1.68–32.5), Po0.01). The incidence of CEP (Po0.001), CV death (Po0.001), stroke (P ¼ 0.001) and myocardial infarction (P ¼ 0.01) increased with increasing LV mass both in patients with and without history of CV disease or diabetes (Figure 1).
a 30 25 20
LV mass
