Articles

Albuminuria in chronic heart failure: prevalence and

prognostic importance
Colette E Jackson, Scott D Solomon, Hertzel C Gerstein, Sofia Zetterstrand, Bertil Olofsson, Eric L Michelson, Christopher B Granger, Karl Swedberg,
Marc A Pfeffer, Salim Yusuf, John J V McMurray for the CHARM Investigators and Committees

Summary
Background Increased excretion of albumin in urine might be a marker of the various pathophysiological changes Lancet 2009; 374: 543–50
that arise in patients with heart failure. Therefore our aim was to assess the prevalence and prognostic value of a spot See Editorial page 501
urinary albumin to creatinine ratio (UACR) in patients with heart failure. See Comment page 506
British Heart Foundation
Methods UACR was measured at baseline and during follow-up of 2310 patients in the Candesartan in Heart failure: Glasgow Cardiovascular
Research Centre, University of
Assessment of Reduction in Mortality and morbidity (CHARM) Programme. The prevalence of microalbuminuria
Glasgow, Glasgow, UK
and macroalbuminuria, and the predictive value of UACR for the primary composite outcome of each CHARM (C E Jackson MB ChB,
study—ie, death from cardiovascular causes or admission to hospital with worsening heart failure—and death from Prof J J V McMurray MD);
any cause were assessed. Brigham and Women’s
Hospital, Boston, MA, USA
(S D Solomon MD,
Findings 1349 (58%) patients had a normal UACR, 704 (30%) had microalbuminuria, and 257 (11%) had Prof M A Pfeffer MD);
macroalbuminuria. The prevalence of increased UACR was similar in patients with reduced and preserved left Department of Medicine and
ventricular ejection fractions. Patients with an increased UACR were older, had more cardiovascular comorbidity, Population Health Research
Institute, McMaster University
worse renal function, and a higher prevalence of diabetes mellitus than did those with normoalbuminuria. However,
and Hamilton Health Sciences,
a high prevalence of increased UACR was still noted among patients without diabetes, hypertension, or renal Hamilton, ON, Canada
dysfunction. Elevated UACR was associated with increased risk of the composite outcome and death even after (Prof H C Gerstein MD,
adjustment for other prognostic variables including renal function, diabetes, and haemoglobin A1c. The adjusted Prof S Yusuf DPhil); AstraZeneca
Research and Development,
hazard ratio (HR) for the composite outcome in patients with microalbuminuria versus normoalbuminuria was Mölndal, Sweden
1·43 (95% CI 1·21–1·69; p<0·0001) and for macroalbuminuria versus normoalbuminuria was 1·75 (1·39–2·20; (S Zetterstrand PhD,
p<0·0001). The adjusted values for death were 1·62 (1·32–1·99; p<0·0001) for microalbuminuria versus normo- B Olofsson PhD); AstraZeneca,
albuminuria, and 1·76 (1·32–2·35; p=0·0001) for macroalbuminuria versus normoalbuminuria. Treatment with Wilmington, DE, USA
(Prof E L Michelson MD); Duke
candesartan did not reduce or prevent the development of excessive excretion of urinary albumin. University Medical Center,
Durham, NC, USA
Interpretation Increased UACR is a powerful and independent predictor of prognosis in heart failure. (Prof C B Granger MD); and
Department of Emergency and
Cardiovascular Medicine,
Funding AstraZeneca. Sahlgrenska Academy,
University of Gothenburg,
Introduction Programme.19,20 We assessed the prevalence of an Gothenburg, Sweden
Increased excretion of albumin in urine is an established increased UACR in these patients at baseline, the (Prof K Swedberg MD)

risk factor for mortality, cardiovascular events, and
adverse renal outcomes in the general population1,2 and
in patients with diabetes,3,4 hypertension,5,6 and other
types of cardiovascular disease.7–9 Screening for increased
albumin excretion is recommended in patients with
diabetes and hypertension to help with risk stratification
and target treatment.10,11 Increased excretion might be a
marker of diffuse vascular injury, systemic inflammation,
activation of the renin-angiotensin system, altered
glomerular haemodynamics, or abnormal tubular
function.12–14 Many, if not all, of these pathophysiological
abnormalities also occur in heart failure.15,16 Measurement
of the urinary albumin to creatinine ratio (UACR) in a
random urine specimen is a convenient method for
detection of increased albumin excretion.17,18 The
prevalence and prognostic importance of elevated UACR
in heart failure, however, are not known.15
We measured UACR in a large subset of patients
enrolled in the Candesartan in Heart failure: Assessment
of Reduction in Mortality and morbidity (CHARM)
characteristics of patients with an increased UACR, and Correspondence to:
Prof John J V McMurray, British
the association between an increased UACR at baseline Heart Foundation Cardiovascular
and subsequent clinical outcomes. Research Centre, University of
Glasgow, 126 University Place,
Methods Glasgow G12 8TA, UK
j.mcmurray@bio.gla.ac.uk
Study design
The design and results of the CHARM trials are described
elsewhere.19,21 Briefly, we studied 7599 patients with New
York Heart Association (NYHA) class II to IV heart
failure with concentrations of serum creatinine of less
than 265 μmol/L (30 mg/L), and serum potassium of less
than 5·5 mmol/L, who were not taking an
angiotensin-receptor blocker, and who had no critical
aortic or mitral stenosis, or recent myocardial infarction,
stroke, or heart surgery. They were divided according to
whether they had left ventricular ejection fraction (LVEF)
greater than 40%; LVEF less than or equal to 40% and
taking an angiotensin-converting enzyme inhibitor; and
LVEF less than or equal to 40% and not taking an
angiotensin-converting enzyme inhibitor because of

www.thelancet.com Vol 374 August 15, 2009 543

 Articles

death from a cardiovascular cause, admission for worsen-

Normoalbuminuria Microalbuminuria Macroalbuminuria p value
(n=1349) (n=704) (n=257) ing heart failure, non-fatal myocardial infarction, non-
fatal stroke, or a coronary revascularisation procedure.
Demographic characteristics
All endpoints were independently adjudicated. The
Age (years) 65 (11·4) 68 (11·4) 68 (10·4) <0·0001
cause of death was considered to be cardiovascular unless
Male sex 870 (64%) 505 (72%) 163 (63%) 0·0596
another clear cause was apparent. Treatment in hospital
Current smoker 183 (14%) 101 (14%) 32 (12%) 0·3727
for worsening heart failure was defined as an unplanned
BMI (kg/m²) 30 (6·3) 29 (6·4) 30 (6·3) 0·0902
admission that was necessitated by heart failure and
Heart failure status
required intravenous diuretics.
NYHA class 0·0055
By design, all patients enrolled in Canada and the USA
II 531 (39%) 254 (36%) 80 (31%) ··
(n=2743) had blood chemistry and haematology measured
III 788 (58%) 428 (61%) 169 (66%) ·· in a central laboratory at baseline, 6 weeks, 14 months,
IV 30 (2%) 22 (3%) 8 (3%) ·· and every year thereafter. Urine was collected, stored,
LVEF (%) 39 (15·5) 39 (16·4) 41 (15·8) 0·1088 and analysed in a central laboratory at baseline, 14 months,
LVEF ≤40% 782 (58%) 423 (60%) 138 (54%) 0·8213 38 months, and the closing study visit. Concentrations of
Physiological measurements serum creatinine were measured spectrophotometrically
HR (beats per min) 72 (12) 72 (12) 74 (13) 0·1235 with the Olympus Chemistry Analyzer (Olympus
SBP (mm Hg) 127 (18) 130 (19) 140 (20) <0·0001 America, Center Valley, PA, USA), urinary albumin with
DBP (mm Hg) 74 (11) 74 (11) 76 (11) 0·1484 a competitive radioimmunoassay Immulite (Diagnostic
Medical history Products, Los Angeles, CA, USA), and urine creatinine
Admission for heart failure 879 (65%) 501 (71%) 183 (71%) 0·0095 with a colourimetric kinetic Jaffe method on the Cobas
Previous myocardial 692 (51%) 369 (52%) 138 (54%) 0·4500 Integra Instrument (Roche Diagnostic Systems,
infarction Branchburg, NJ, USA).7
Angina pectoris 822 (61%) 421 (60%) 152 (59%) 0·5097
PCI 242 (18%) 136 (19%) 47 (18%) 0·5685 Statistical analysis
CABG 371 (28%) 239 (34%) 97 (38%) 0·0001 These analyses were restricted to the North American
Pacemaker 134 (10%) 81 (12%) 25 (10%) 0·5219 participants for whom UACRs were available. Macro-
Diabetes mellitus 387 (29%) 307 (44%) 169 (66%) <0·0001 albuminuria was defined as UACR greater than
Hypertension 823 (61%) 481 (68%) 210 (82%) <0·0001 25 mg/mmol in men and women, and microalbuminuria
Atrial fibrillation 353 (26%) 258 (37%) 87 (34%) <0·0001 as UACR 2·5–25·0 mg/mmol in men and
Stroke 125 (9%) 76 (11%) 43 (17%) 0·0031 3·5–25·0 mg/mmol in women.22 Cox proportional
Laboratory measurements hazards models were used to analyse the prospective
Serum creatinine (μmol/L) 97 (33·3) 111 (39·4) 123 (47·2) <0·0001 association between UACR and death from a
eGFR (mL/min/1·73m²) 81·2 (29·3) 70·6 (28·9) 65·1 (31·8) <0·0001 cardiovascular cause or admission for worsening heart
<60 275 (20%) 244 (35%) 119 (46%) <0·0001 failure, all-cause death, and admission to hospital for
<30 17 (1%) 20 (3%) 19 (7%) <0·0001 worsening heart failure. The association between UACR
Potassium (mmol/L) 4·4 (0·4) 4·4 (0·5) 4·5 (0·5) 0·0901 and these outcomes was adjusted for 33 baseline variables
Haemoglobin (mg/L) 85 (9) 84 (11) 82 (11) 0·0001 as described previously:19,20 randomly assigned treatment
Haematocrit (%) 41 (4·4) 41 (5·2) 40 (5·5) 0·0027 (candesartan vs placebo), sex (men vs women), NYHA
HbA1c (%) 6·8 (1·3) 7·4 (1·7) 7·9 (1·8) <0·0001 class (III/IV vs II), smoking habit (current vs none or
History of diabetes 8·0 (1·6) 8·5 (1·7) 8·6 (1·8) <0·0001 past), age, LVEF, body-mass index, systolic blood pressure,
No history of diabetes 6·2 (0·7) 6·4 (0·8) 6·6 (0·7) <0·0001 diastolic blood pressure, heart rate, history (admission for
(Continues on next page) heart failure, myocardial infarction, angina pectoris,
stroke, hypertension, diabetes mellitus, atrial fibrillation,
cancer, coronary artery bypass surgery, percutaneous
intolerance. Within each of the component trials, patients coronary revascularisation, implanted cardioverter
were randomly allocated to candesartan (up to 32 mg defibrillator, or pacemaker), and baseline treatment
once a day) or matching placebo. Median follow-up of the (diuretic, digitalis, β blocker, angiotensin-
entire cohort was 37·7 months. converting enzyme inhibitor, calcium-channel blocker,
The primary outcome of the entire programme was other vasodilator, antiarrhythmic drug, lipid-lowering
death from any cause and the primary composite outcome drug, anticoagulant, aspirin, and other antiplatelet).
for the three component trials was death from a UACR was added to this model, as a categorical and
cardiovascular cause or admission for worsening heart continuous variable (34-variable model). Concentrations
failure. Other prespecified outcomes were death from any of serum creatinine and haemoglobin A1c (HbA1c) were
cause or admission for heart failure; death from a cardio- then added simultaneously to the model with UACR
vascular cause, admission for worsening heart failure, (36-variable model) to assess risk associated with increased
non-fatal myocardial infarction, or non-fatal stroke; and UACR, independently of renal dysfunction and

544 www.thelancet.com Vol 374 August 15, 2009

 Articles

dysglycaemia.23,24 An alternative 36-variable model was

Normoalbuminuria Microalbuminuria Macroalbuminuria p value
also run with an estimated glomerular filtration rate (n=1349) (n=704) (n=257)
(eGFR) with the four-variable equation for modification
(Continued from previous page)
of diet in renal disease instead of concentration of serum
creatinine.25 Drugs at randomisation
ACE inhibitor 585 (43%) 299 (42%) 107 (42%) 0·5657
Role of the funding source β blocker 725 (54%) 366 (52%) 138 (54%) 0·6248
The collection and analysis of urine samples for albumin Diuretic 1124 (83%) 626 (89%) 234 (91%) <0·0001
excretion was a preplanned and investigator-originated Nitrate 384 (28%) 243 (35%) 88 (34%) 0·0031
substudy. The measurement of urinary albumin Spironolactone 202 (15%) 108 (15%) 25 (10%) 0·2098
concentrations was paid for by the sponsor and done at Digoxin 700 (52%) 385 (55%) 133 (52%) 0·4600
McMaster University and Hamilton Health Sciences, Calcium-channel blocker 326 (24%) 180 (26%) 100 (39%) 0·0003
Hamilton, ON, Canada. The statistical analyses were Lipid-lowering drug 620 (46%) 337 (48%) 134 (52%) 0·0925
done by SZ who is an employee of the sponsor. Oral anticoagulant 396 (29%) 245 (35%) 79 (31%) 0·0594

Data are mean (SD) or number (%). BMI=body-mass index. NYHA=New York Heart Association. LVEF=left ventricular
Results ejection fraction. HR=heart rate. SBP=systolic blood pressure. DBP=diastolic blood pressure. PCI=percutaneous
UACR was measured in 2310 (84%) of 2743 North coronary intervention. CABG=coronary artery bypass grafting. eGFR=estimated glomerular filtration rate.
American patients in CHARM. Table 1 shows the HbA1c=haemoglobin A1c. ACE=angiotensin-converting enzyme.

baseline characteristics of the three groups: 58% had a
normal UACR, 30% had microalbuminuria, and 11%
had macroalbuminuria. Patients with an increased
UACR were older, had a higher systolic blood pressure,
and were more likely to have a history of hypertension
and diabetes (especially those with macroalbuminuria)
than were those with normal UACR. A history of stroke,
coronary heart disease, and atrial fibrillation was also
more common in patients with an elevated UACR. HbA1c
level was higher in patients with an elevated UACR (even
in those without diabetes), and renal dysfunction
(defined as eGFR <60 mL/min/1·73 m²) was more
common in these patients than in those with normal
UACR. Patients with an increased UACR were also more
likely to have been admitted for heart failure, and a
higher proportion had NYHA functional class III or IV
symptoms at the time of randomisation. The use of
angiotensin-converting enzyme inhibitors, β blockers,
and digoxin was similar in all three UACR groups
although the proportion of patients being treated with
spironolactone was smallest in the group with macro-
albuminuria. Nitrates were used more commonly in
patients with an elevated UACR, and use of
Table 1: Characteristics of patients with and without an elevated urinary albumin to creatinine ratio
calcium-channel blockers was more common in those
with macroalbuminuria (table 1).
Of 1447 patients without diabetes, 397 (27%) had micro-
albuminuria and 88 (6%) had macroalbuminuria.
307 (36%) of 863 with diabetes had microalbuminuria
and 169 (20%) had macroalbuminuria.
Among 1714 patients with a systolic blood pressure of
140 mm Hg or less, or a diastolic blood pressure of
90 mm Hg or less, 508 (30%) had microalbuminuria and
146 (9%) had macroalbuminuria. Of 596 patients with a
systolic blood pressure of more than 140 mm Hg or a
diastolic blood pressure of more than 90 mm Hg,
196 (33%) had microalbuminuria and 111 (19%) had
macroalbuminuria.
Of 796 patients without a history of hypertension,
223 (28%) had microalbuminuria and 47 (6%) had
macroalbuminuria. In 1514 patients with a history of
hypertension, 481 (32%) had microalbuminuria and
210 (14%) had macroalbuminuria. Of 755 patients
without either diabetes or a history of hypertension,

Normoalbuminuria Microalbuminuria Hazard ratio* Macroalbuminuria Hazard ratio*

(n=1349) (n=704) (95% CI) (n=257) (95% CI)
Cardiovascular death or admission for heart failure 397 (28%) 306 (43%) 1·74 (1·50–2·02) 135 (52%) 2·36 (1·94–2·86)
Cardiovascular death† 196 (13%) 188 (27%) ·· 65 (26%) ··
Admission for heart failure† 292 (22%) 213 (32%) ·· 102 (41%) ··
All-cause mortality or admission for heart failure 437 (31%) 332 (46%) 1·71 (1·48–1·97) 146 (55%) 2·32 (1·92–2·79)
Cardiovascular death or admission for heart failure, non-fatal myocardial 440 (31%) 328 (46%) 1·69 (1·47–1·95) 146 (56%) 2·35 (1·95–2·84)
infarction, or non-fatal stroke
Cardiovascular death, admission for heart failure, non-fatal myocardial 488 (35%) 345 (48%) 1·59 (1·38–1·82) 153 (59%) 2·22 (1·85–2·66)
infarction, non-fatal stroke, or coronary revascularisation procedure
All-cause mortality 249 (17%) 231 (31%) 1·97 (1·65–2·36) 87 (33%) 2·07 (1·63–2·65)

Data are number (%); percentages are Kaplan-Meier failure rates at 3 years. *Unadjusted hazard ratio (compared with normoalbuminuria). †Any occurrence of these outcomes.

Table 2: Clinical outcomes according to baseline urinary albumin to creatinine ratio

www.thelancet.com Vol 374 August 15, 2009 545

 Articles

macroalbuminuria. In 967 patients with heart failure and

A a preserved LVEF (>40%), 281 (29%) had microalbumin-
0·7 Macroalbuminuria
Microalbuminuria
uria and 119 (12%) had macroalbuminuria.
Normoalbuminuria 405 (31%) of 1319 patients who were not taking an
0·6
Estimated cumulative distribution function

angiotensin-converting enzyme inhibitor at baseline

had microalbuminuria and 150 (11%) had macro-
0·5
albuminuria. 299 (30%) of 991 patients taking an
0·4
angiotensin-converting enzyme inhibitor at baseline
had microalbuminuria and 107 (11%) had macro-
0·3
albuminuria.
The unadjusted risk of adverse cardiovascular
0·2 outcomes was increased in patients with an elevated
UACR (table 2). Fewer patients with a normal UACR
0·1 had the primary composite outcome of death from a
cardiovascular cause or admission for heart failure than
0 did those with microalbuminuria or macroalbuminuria
Number at risk (figure 1; table 2). Elevated UACR was associated with
Normoalbuminuria 1346 1246 1168 1099 1013 817 411 an increased risk of each of the components of this
Microalbuminuria 703 592 547 487 434 326 148
Macroalbuminuria 256 209 174 153 136 100 45
outcome. This increased risk was also noted for the
expanded secondary cardiovascular outcomes that
B included myocardial infarction, stroke, and coronary
0·45 revascularisation (table 2). Notably, the proportions of
patients admitted to hospital with heart failure were
0·40 increased substantially in the elevated UACR
categories—with the highest increase in patients with
0·35 macroalbuminuria (table 2). The proportion of patients
Estimated cumulative distribution function

who died from any cause also increased with increase in

0·30 UACR (figure 2; table 2).
The albuminuria category was an independent,
0·25 significant, predictor of the primary outcome, all-cause
mortality, and admission for heart failure when added
0·20
to the Cox regression analysis with 33 baseline
demographic and clinical factors as covariates (table 3).
0·15
For each of these outcomes, patients with microalbumin-
uria or macroalbuminuria had a 40–80% increase in
0·10
adjusted risk. UACR was also an independent predictor
of outcome when analysed as a continuous variable. An
0·05
increase in UACR of 100 mg/mmol was associated with
0
about a 10% increase in adjusted risk for all three
0 0·5 1·0 1·5 2·0 2·5 3·0 outcomes.
Number at risk
Time (years) Creatinine was a significant independent predictor of
Normoalbuminuria 1348 1312 1270 1234 1170 964 492 all three outcomes whereas the predictive value of HbA1c
Microalbuminuria 704 657 632 589 542 421 192
Macroalbuminuria 256 242 229 211 195 150 64
was not always significant for the primary outcome in
the multivariable analysis that included UACR, HbA1c,
Figure 1: Cardiovascular death or admission for chronic heart failure (A), and all-cause mortality (B), stratified and creatinine (table 4). However, UACR remained an
by albuminuria status independent predictor when both creatinine and HbA1c
were added to the model, with little reduction in the
212 (28%) had microalbuminuria and 43 (6%) had hazard ratios (HRs) related to microalbuminuria and
macroalbuminuria. macroalbuminuria for any of the three outcomes (table 4).
372 (27%) of 1375 patients with an eGFR greater than Similar results were obtained when this 36-covariate
or equal to 60 mL/min/1·73m² had microalbuminuria model was re-run with eGFR instead of creatinine (data
and 108 (8%) had macroalbuminuria. Of 638 patients not shown).
with renal impairment (eGFR <60 mL/min/1·73m²), The risk associated with excretion of albumin in urine,
244 (38%) had microalbuminuria and 119 (19%) had assessed as a continuous or categorical variable, was
macroalbuminuria. similar in patients with low LVEF and in those with
423 (31%) of 1343 patients with heart failure and a low preserved LVEF. HR for the primary composite outcome
LVEF (≤40%) had microalbuminuria and 138 (10%) had in the categorical analysis was 1·84 (95% CI 1·39–2·43)

546 www.thelancet.com Vol 374 August 15, 2009

 Articles

60 Cardiovascular death or admission for heart failure

Death from any cause
50
Clinical outcomes (%)

40

30

20

10

0
1 2 3 4 5 6 7 8 9 10
Decile of UACR
0–0·40 0·40–0·66 0·66–0·94 0·94–1·32 1·32–1·94 1·94–3·01 3·01–5·04 5·04–10·56 10·56–29·53 >29·53
(n=231) (n=231) (n=231) (n=231) (n=231) (n=232) (n=230) (n=231) (n=231) (n=231)

Figure 2: Clinical outcomes according to decile of urinary albumin to creatinine ratio (UACR)

for macroalbuminuria versus normoalbuminuria and Studies of Left Ventricular Dysfunction (SOLVD) had a
1·61 (1·34–1·95) for microalbuminuria versus normo- urine dipstick test for protein at baseline.26 Of 5487 (81%
albuminuria in patients with a low LVEF compared with of total) tested, 177 (3%) had proteinuria. These patients
2·03 (1·45–2·85) for macroalbuminuria versus normo- had a higher blood pressure and a higher prevalence of
albuminuria, and 1·31 (0·99–1·74) for microalbumin- diabetes mellitus, and also greater left ventricular
uria versus normoalbuminuria in those with a preserved systolic dysfunction and more symptomatic heart
LVEF (p=0·1566 for interaction between LVEF and failure than did those without proteinuria. The results
macroalbuminuria, and p=0·8280 for interaction of dipstick urine testing for proteinuria were also
between LVEF and microalbuminuria). HR per unit reported in a Canadian subset (n=583) of 2231 patients
UACR (100 mg/mmol) was 1·10 (0·99–1·21) in patients
with low LVEF and 1·12 (1·04–1·21) in those with
preserved LVEF (p=0·6283 for interaction). Hazard ratio p value
(95% CI)
The risk related to excretion of albumin in urine was
also evident in the subset of patients with a normal Cardiovascular death or admission for heart failure
UACR at baseline (and of at least the same magnitude as Albuminuria category*
in the overall patient population; figure 2). HR for the Microalbuminuria vs 1·50 (1·28–1·75) <0·0001
primary outcome for each unit increase in UACR was normoalbuminuria

1·15 (95% CI 1·00–1·32; p=0·0565). Urine samples were Macroalbuminuria vs 1·88 (1·53–2·33) <0·0001
normoalbuminuria
not obtained at follow-up in a large proportion of patients
UACR continuous (100 mg/mmol)† 1·10 (1·04–1·17) 0·0016
(table 5). There was no evidence to suggest that treatment
All-cause mortality
with candesartan prevented the development or reduced
Albuminuria category*
the excessive excretion of albumin in urine.
Microalbuminuria vs 1·63 (1·35–1·97) <0·0001
normoalbuminuria
Discussion Macroalbuminuria vs 1·77 (1·36–2·31) <0·0001
The prevalence of elevated UACR in patients with heart normoalbuminuria
failure was high and was associated with a substantially UACR continuous (100 mg/mmol)† 1·10 (1·02–1·19) 0·0153
increased risk of adverse clinical outcomes, including Admission for heart failure
death. Even after adjustment for other risk factors in a Albuminuria category*
multivariable model, microalbuminuria or macro- Microalbuminuria vs 1·41 (1·17–1·69) 0·0003
albuminuria remained strong independent predictors, normoalbuminuria
with a 60–80% adjusted increase in the risk of death and Macroalbuminuria vs 1·88 (1·47–2·40) <0·0001
a 30–70% increase in the adjusted risk of admission for normoalbuminuria
heart failure. UACR continuous (100 mg/mmol)† 1·11 (1·04–1·19) 0·0024
Little is known about excretion of albumin in urine in *UACR analysed as a categorical variable (microalbuminuria or macroalbuminuria
patients with heart failure and nothing is known about vs normoalbuminuria). †UACR analysed as a continuous variable.
its prognostic importance. Few data are available for the
Table 3: Multivariable analysis of urinary albumin to creatinine ratio
prevalence of increased excretion of albumin in the
(UACR) added to 33 baseline covariates by endpoint
urine of patients with heart failure. Most patients in the

www.thelancet.com Vol 374 August 15, 2009 547

 Articles
Placebo at baseline (n=1142)
Normoalbuminuria (n=656)
Microalbuminuria (n=355)
Macroalbuminuria (n=131)
Candesartan at baseline (n=1168)
Normoalbuminuria (n=693)
Microalbuminuria (n=349)
Macroalbuminuria (n=126)
Data are number (%).
Table 5: Urinary albumin to creatinine ratio shifts between baseline and study follow-up visit at 14 months
548
with left ventricular systolic dysfunction after myocardial subset had a trace of proteinuria and 6% had greater
infarction who were enrolled in the Survival and than trace proteinuria. Patients with proteinuria had a
Ventricular Enlargement trial (SAVE).27 15% of this higher baseline creatinine and Killip class than did
those without proteinuria. They also had a lower LVEF
Hazard ratio p value but had a similar prevalence of diabetes to those without
(95% CI)
proteinuria.
Cardiovascular death or admission for heart failure By contrast with these two studies,26,27 UACR was
UACR categorical* measured in one other study in which 29 (30%) of 96
Microalbuminuria vs normoalbuminuria 1·43 (1·21–1·69) <0·0001 outpatients (mean age 69 years; 22% women) with
Macroalbuminuria vs normoalbuminuria 1·75 (1·39–2·20) <0·0001 predominantly NYHA class III heart failure had micro-
Creatinine (10 μmol/L) 1·04 (1·02–1·06) <0·0001 albuminuria, and 5 (5%) had macroalbuminuria.15
HbA1c (%) 1·04 (0·99–1·11) 0·1466 Although differences between patients with and without
UACR continuous† microalbuminuria were noted, they were not significant,
UACR (100 mg/mmol) 1·07 (1·00–1·14) 0·0414 probably because there were few patients in each group
Creatinine (10 μmol/L) 1·05 (1·03–1·07) <0·0001 (and patients with macroalbuminuria were not described
HbA1c (%) 1·06 (1·00–1·12) 0·0441 further).
All-cause mortality In our study, microalbuminuria was much more
UACR categorical* common than was macroalbuminuria. Patients with an
Microalbuminuria vs normoalbuminuria 1·62 (1·32–1·99) <0·0001 increased UACR were older and had more cardiovascular
Macroalbuminuria vs normoalbuminuria 1·76 (1·32–2·35) 0·0001 problems, dysglycaemia, and renal impairment, and also
Creatinine (10 μmol/L) 1·03 (1·01–1·05) 0·0171 had worse symptoms of heart failure than did those
HbA1c (%) 1·03 (0·96–1·11) 0·4429 without increased UACR. LVEF did not differ between
UACR continuous† those with and without an increased UACR, and the
UACR (100 mg/mmol) 1·08 (1·00–1·17) 0·0598
corollary was also true—ie, the prevalence of an increased
Creatinine (10 μmol/L) 1·04 (1·01–1·06) 0·0025
UACR was similar in patients with reduced and
HbA1c (%) 1·05 (0·98–1·13) 0·1821
preserved LVEF heart failure. However, a third of patients
without diabetes had microalbuminuria or macro-
Admission for heart failure
albuminuria, and more than a third of those without
UACR categorical*
hypertension or renal impairment also had an elevated
Microalbuminuria vs normoalbuminuria 1·31 (1·07–1·59) 0·0077
UACR. Consequently, increased excretion of albumin in
Macroalbuminuria vs normoalbuminuria 1·67 (1·28–2·17) 0·0002
urine in patients with heart failure cannot be wholly
Creatinine (10 μmol/L) 1·06 (1·03–1·08) <0·0001
explained by concomitant diabetes, hypertension, or
HbA1c (%) 1·04 (0·98–1·11) 0·2223
renal dysfunction. Notably, the prevalence of micro-
UACR continuous†
albuminuria in patients without diabetes was three-fold
UACR (100 mg/mmol) 1·07 (0·99–1·15) 0·0753
greater than that in age-matched individuals in the
Creatinine (10 μmol/L) 1·06 (1·04–1·09) <0·0001
general population; the prevalence of microalbuminuria
HbA1c (%) 1·06 (0·99–1·13) 0·0998
in Dutch individuals aged 60–74 years was 10·4%
HbA1c=haemoglobin A1c. *UACR analysed as a categorical variable (95% CI 9·8–11·0).15
(microalbuminuria or macroalbuminuria vs normoalbuminuria). †UACR analysed The mechanism underlying albuminuria in patients
as a continuous variable.
without these conditions is not known. Renal venous
Table 4: Multivariable analysis of urinary albumin to creatinine ratio congestion caused proteinuria in dogs, and urinary
(UACR), serum creatinine concentration, and haemoglobin A1c added to albumin excretion was inversely related to renal blood
33 baseline covariates by endpoint flow in patients with heart failure.28–30 Increased albumin
excretion might therefore have a haemodynamic basis in
Normoalbuminuria Microalbuminuria Macroalbuminuria Missing heart failure, particularly when renal venous congestion
is associated with reduced renal blood flow. All of these
264 (40%) 88 (13%) 9 (1%) 295 (45%)
renal changes are associated with worse outcomes in
57 (16%) 118 (33%) 25 (7%) 155 (44%)
patients with heart failure.28–30
4 (3%) 27 (21%) 44 (34%) 56 (43%)
Even after the differences between those with and
without an elevated UACR are accounted for in the
multivariable analyses, increased UACR remained a
278 (40%) 72 (10%) 8 (1%) 335 (48%)
significant independent predictor of the primary
88 (25%) 116 (33%) 18 (5%) 127 (36%)
composite outcome of cardiovascular death or admission
5 (4%) 28 (22%) 38 (30%) 55 (44%)
for worsening heart failure, and all-cause mortality. Of
note, increased urinary albumin excretion was also an
independent predictor of the disease-specific non-fatal
outcome of admission for worsening heart failure. The
www.thelancet.com Vol 374 August 15, 2009

small proportion of patients with dipstick proteinuria in
SOLVD also had an increased risk of death and of
admission for heart failure.26 Dipstick proteinuria was
also a significant independent predictor of mortality in
SAVE.27
The hazard related to increased urinary albumin excre-
tion was independent of renal dysfunction, as assessed
by both serum creatinine concentration and eGFR, and
also dysglycaemia, with adjustment for history of
diabetes and HbA1c. The risk related to increased
albumin excretion in urine was apparent in patients
with low LVEF and also those with preserved LVEF
heart failure.
Another important finding, lending support to recent
findings in patients with stable atherosclerotic disease,
is that even an increasing UACR within the normo-
albuminuric range was associated with an increased
hazard of adverse clinical outcomes, showing that
urinary albumin excretion represents a continuous
measure of risk.8
Candesartan did not reduce excessive albumin
excretion or prevent its development although the
interpretation of this analysis is uncertain with the high
proportion of missing follow-up urine samples and is
probably too soon to conclude that angiotensin-receptor
blockers definitely have no effect on urinary albumin
excretion in patients with heart failure. These drugs
reduce albumin excretion in patients with type 2 diabetes
mellitus, hypertension, substantial renal impairment,
and marked proteinuria. Whether the pathogenesis of
increased albumin excretion in heart failure is the same
and whether an angiotensin-receptor blocker should
reduce albumin excretion in patients with heart failure
are not known.
One limitation of our study is that patients were
selected from those enrolled in a clinical trial, and,
particularly, those with severe renal dysfunction were
not included. However, the proportions of patients with
microalbuminuria and macroalbuminuria in our study
were similar to those reported in the only other study
done in patients with heart failure.15 Newer methods for
measurement of renal function, such as levels of
cystatin C, and other important prognostic markers such
as natriuretic peptides and troponin were not measured
in CHARM.
Because UACR is a simple, readily available clinical
test that is widely used in primary and secondary care, it
might be of value in risk stratification of patients with
heart failure. However, whether UACR adds incremental
prognostic information to other new prognostic
biomarkers, particularly natriuretic peptides, which are
also increasingly and widely used in clinical practice,
needs to be assessed. Of potential interest to physicians
and patients is whether therapeutic reduction in albumin
excretion, which did not occur with candesartan, might
be useful in the prediction of improvement in clinical
outcomes.
www.thelancet.com Vol 374 August 15, 2009
Articles
Contributors
MAP, KS, CBG, JJVM, SY, BO, and ELM designed the CHARM
Programme and substudies. HCG and SDS also contributed to the
design and interpretation of the present study. SZ and BO did the
statistical analyses. All authors contributed to the interpretation of the
data. CEJ and JJVM wrote the draft of the report, and all authors
contributed to its revision. JJVM takes responsibility for the report.
Conflicts of interest
MAP, KS, HCG, CBG, JJVM, SDS, and SY have received research
funding, and lecture and consulting fees from AstraZeneca. ELM, BO,
and SZ are employees of AstraZeneca. CEJ declares that she has no
conflicts of interest.
Acknowledgments
The CHARM Programme was funded by AstraZeneca.
References
1 Arnlöv J, Evans JC, Meigs JB, et al. Low-grade albuminuria and
incidence of cardiovascular disease events in nonhypertensive and
nondiabetic individuals: the Framingham Heart Study. Circulation
2005; 112: 969–75.
2 Diercks GF, van Boven AJ, Hillege HL, et al. Microalbuminuria is
independently associated with ischaemic electrocardiographic
abnormalities in a large non-diabetic population. The PREVEND
(Prevention of REnal and Vascular ENdstage Disease) study.
Eur Heart J 2000; 21: 1922–27.
3 Mogensen CE. Microalbuminuria predicts clinical proteinuria and
early mortality in maturity-onset diabetes. N Engl J Med 1984;
310: 356–60.
4 Deckert T, Yokoyama H, Mathiesen E, et al. Cohort study of predictive
value of urinary albumin excretion for atherosclerotic vascular disease
in patients with insulin dependent diabetes. BMJ 1996; 312: 871–74.
5 Agewall S, Wikstrand J, Ljungman S, Fagerberg B. Usefulness of
microalbuminuria in predicting cardiovascular mortality in treated
hypertensive men with and without diabetes mellitus. Risk Factor
Intervention Study Group. Am J Cardiol 1997; 80: 164–69.
6 Wachtell K, Ibsen H, Olsen MH, et al. Albuminuria and
cardiovascular risk in hypertensive patients with left ventricular
hypertrophy: the LIFE study. Ann Intern Med 2003; 139: 901–06.
7 Gerstein HC, Mann JF, Yi Q, et al; HOPE Study Investigators.
Albuminuria and risk of cardiovascular events, death, and heart
failure in diabetic and nondiabetic individuals. JAMA 2001;
286: 421–26.
8 Solomon SD, Lin J, Solomon CG, et al. Influence of albuminuria on
cardiovascular risk in patients with stable coronary artery disease.
Circulation 2007; 116: 2687–93.
9 Brantsma AH, Bakker SJ, Hillege HL, et al. Cardiovascular and renal
outcome in subjects with K/DOQI stage 1-3 chronic kidney disease:
the importance of urinary albumin excretion. Nephrol Dial Transplant
2008; 23: 3851–58.
10 Mancia G, De Backer G, Dominiczak A, et al. 2007 Guidelines for the
management of arterial hypertension: the Task Force for the
Management of Arterial Hypertension of the European Society of
Hypertension (ESH) and of the European Society of Cardiology
(ESC). J Hypertens 2007; 25: 1105–87.
11 American Diabetes Association. Standards of medical care in
diabetes–2008. Diabetes Care 2008; 31 (suppl 1): S12–54.
12 Deckert T, Feldt-Rasmussen B, Borch-Johnsen K, et al. Albuminuria
reflects widespread vascular damage. The Steno hypothesis.
Diabetologia 1989; 32: 219–26.
13 Chugh A, Bakris GL. Microalbuminuria: what is it? Why is it
important? What should be done about it? An update.
J Clin Hypertens (Greenwich) 2007; 9: 196–200.
14 Weir MR. Microalbuminuria and cardiovascular disease.
Clin J Am Soc Nephrol 2007; 2: 581–90.
15 van de Wal RM, Asselbergs FW, Plokker HW, et al. High prevalence
of microalbuminuria in chronic heart failure patients. J Card Fail
2005; 11: 602–06.
16 Silverberg D, Wexler D, Blum M, et al. The association between
congestive heart failure and chronic renal disease.
Curr Opin Nephrol Hypertens 2004; 13: 163–70.
17 Jensen JS, Clausen P, Borch-Johnsen K, et al. Detecting micro-
albuminuria by urinary albumin/creatinine concentration ratio.
Nephrol Dial Transplant 1997; 12 (suppl 2): 6–9.
549
 Articles
18 Keane WF, Eknoyan G. Proteinuria, albuminuria, risk, assessment,
detection, elimination (PARADE): a position paper of the National
Kidney Foundation. Am J Kidney Dis 1999; 33: 1004–10.
19 Pfeffer MA, Swedberg K, Granger CB, et al. Effects of candesartan
on mortality and morbidity in patients with chronic heart failure:
the CHARM-Overall programme. Lancet 2003; 362: 759–66.
20 Pocock SJ, Wang D, Pfeffer MA, et al. Predictors of mortality and
morbidity in patients with chronic heart failure. Eur Heart J 2006;
27: 65–75.
21 Swedberg K, Pfeffer M, Granger C, et al. Candesartan in heart
failure--assessment of reduction in mortality and morbidity
(CHARM): rationale and design. Charm-Programme Investigators.
J Card Fail 1999; 5: 276–82.
22 IDF Clinical Guidelines Task Force. Global Guideline for Type 2
Diabetes: recommendations for standard, comprehensive, and
minimal care. Diabet Med 2006; 23: 579–93.
23 Hillege HL, Nitsch D, Pfeffer MA, et al. Renal function as a
predictor of outcome in a broad spectrum of patients with heart
failure. Circulation 2006; 113: 671–78.
24 Gerstein HC, Swedberg K, Carlsson J, et al. The hemoglobin A1c
level as a progressive risk factor for cardiovascular death,
hospitalization for heart failure, or death in patients with chronic
heart failure: an analysis of the Candesartan in Heart failure:
Assessment of Reduction in Mortality and Morbidity (CHARM)
program. Arch Intern Med 2008; 168: 1699–704.
550
25 Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more
accurate method to estimate glomerular filtration rate from serum
creatinine: a new prediction equation. Modification of Diet in Renal
Disease Study Group. Ann Intern Med 1999; 130: 461–70.
26 Capes SE, Gerstein HC, Negassa A, Yusuf S. Enalapril prevents
clinical proteinuria in diabetic patients with low ejection fraction.
Diabetes Care 2000; 23: 377–80.
27 Jose P, Tomson C, Skali H, et al. Influence of proteinuria on
cardiovascular risk and response to angiotensin-converting enzyme
inhibition after myocardial infarction. J Am Coll Cardiol 2006;
47: 1725–27.
28 Wegria R, Capeci NE, Blumenthal MR, et al. The pathogenesis of
proteinuria in the acutely congested kidney. J Clin Invest 1955;
34: 737–43.
29 Smilde TD, Damman K, van der Harst P, et al. Differential
associations between renal function and “modifiable” risk factors
in patients with chronic heart failure. Clin Res Cardiol 2009;
98: 121–29.
30 Damman K, van Deursen VM, Navis G, et al. Increased central
venous pressure is associated with impaired renal function and
mortality in a broad spectrum of patients with cardiovascular
disease. J Am Coll Cardiol 2009; 53: 582–88.
www.thelancet.com Vol 374 August 15, 2009