Selective vitamin D receptor activation with paricalcitol for reduction of albuminuria in patients with type 2 diabetes (VITAL study): a randomised controlled trial.

Abstract

Despite treatment with renin–angiotensin–aldosterone system (RAAS) inhibitors, patients with diabetes have increased risk of progressive renal failure that correlates with albuminuria. We aimed to assess whether paricalcitol could be used to reduce albuminuria in patients with diabetic nephropathy.
In this multinational, placebo-controlled, double-blind trial, we enrolled patients with type 2 diabetes and albuminuria who were receiving angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. Patients were assigned (1:1:1) by computer-generated randomisation sequence to receive 24 weeks’ treatment with placebo,1 μg/day paricalcitol, or 2 μg/day paricalcitol. The primary endpoint was the percentage change in geometric mean urinary albumin-to-creatinine ratio (UACR) from baseline to last measurement during treatment for the combined paricalcitol groups versus the placebo group. Analysis was by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00421733.
Between February, 2007, and October, 2008, 281 patients were enrolled and assigned to receive placebo(n=93), 1 μg paricalcitol (n=93), or 2 μg paricalcitol (n=95); 88 patients on placebo, 92 on 1 μg paricalcitol, and 92 on2 μg paricalcitol received at least one dose of study drug, and had UACR data at baseline and at least one timepoint during treatment, and so were included in the primary analysis. Change in UACR was: –3% (from 61 to 60 mg/mmol;95% CI –16 to 13) in the placebo group; –16% (from 62 to 51 mg/mmol; –24 to –9) in the combined paricalcitol groups, with a between-group difference versus placebo of –15% (95% CI –28 to 1; p=0.071); –14% (from 63 to 54 mg/mmol; –24 to –1) in the 1 μg paricalcitol group, with a between-group difference versus placebo of –11%(95% CI –27 to 8; p=0.23); and –20% (from 61 to 49 mg/mmol; –30 to –8) in the 2 μg paricalcitol group, with a between-group difference versus placebo of –18% (95% CI –32 to 0; p=0.053). Patients on 2 μg paricalcitol showed a nearly, sustained reduction in UACR, ranging from –18% to –28% (p=0.014 vs placebo). Incidence of hypercalcaemia,adverse events, and serious adverse events was similar between groups receiving paricalcitol versus placebo.
Addition of 2 μg/day paricalcitol to RAAS inhibition safely lowers residual albuminuria in patients with diabetic nephropathy, and could be a novel approach to lower residual renal risk in diabetes.
Abbott.

Full text

Page 1

Articles
www.thelancet.com Vol 376 November 6, 2010 1543
Lancet 2010; 376: 1543–51
Published Online
November 4, 2010
DOI:10.1016/ S0140-
6736(10)61032-X
See Comment page 1521
University Medical Centre
Groningen, University of
Groningen, Groningen,
Netherlands
(Prof D de Zeeuw MD); Indiana
University School of Medicine,
Indianapolis, IN, USA
(Prof R Agarwal MD); Abbott
Laboratories, Chicago, IL, USA
(M Amdahl MS, P Audhya MD,
T Garimella PhD, Y Pritchett PhD,
D Andress MD); Washington
University School of Medicine,
St Louis, MO, USA
(Prof D Coyne MD);
Rigshospitalet, Copenhagen
and Aarhus University, Aarhus,
Denmark
(Prof H-H Parving MD); Mario
Negri Institute for
Pharmacological Research,
Bergamo, Italy
(Prof G Remuzzi MD); and
University of Heidelberg,
Heidelberg, Germany
(Prof E Ritz MD)
Correspondence to:
Prof Dick de Zeeuw, Department
of Clinical Pharmacology,
University Medical Centre
Groningen, PO Box 196,
9700 AD Groningen,
Netherlands
d.de.zeeuw@med.umcg.nl
Selective vitamin D receptor activation with paricalcitol for
reduction of albuminuria in patients with type 2 diabetes
(VITAL study): a randomised controlled trial
Dick de Zeeuw, Rajiv Agarwal, Michael Amdahl, Paul Audhya, Daniel Coyne, Tushar Garimella, Hans-Henrik Parving, Yili Pritchett,
Giuseppe Remuzzi, Eberhard Ritz, Dennis Andress
Summary
Background Despite treatment with renin–angiotensin–aldosterone system (RAAS) inhibitors, patients with diabetes
have increased risk of progressive renal failure that correlates with albuminuria. We aimed to assess whether
paricalcitol could be used to reduce albuminuria in patients with diabetic nephropathy.
Methods In this multinational, placebo-controlled, double-blind trial, we enrolled patients with type 2 diabetes and
albuminuria who were receiving angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. Patients
were assigned (1:1:1) by computer-generated randomisation sequence to receive 24 weeks’ treatment with placebo,
1 μg/day paricalcitol, or 2 μg/day paricalcitol. The primary endpoint was the percentage change in geometric mean
urinary albumin-to-creatinine ratio (UACR) from baseline to last measurement during treatment for the combined
paricalcitol groups versus the placebo group. Analysis was by intention to treat. This trial is registered with
ClinicalTrials.gov, number NCT00421733.
Findings Between February, 2007, and October, 2008, 281 patients were enrolled and assigned to receive placebo
(n=93), 1 μg paricalcitol (n=93), or 2 μg paricalcitol (n=95); 88 patients on placebo, 92 on 1 μg paricalcitol, and 92 on
2 μg paricalcitol received at least one dose of study drug, and had UACR data at baseline and at least one timepoint
during treatment, and so were included in the primary analysis. Change in UACR was: –3% (from 61 to 60 mg/mmol;
95% CI –16 to 13) in the placebo group; –16% (from 62 to 51 mg/mmol; –24 to –9) in the combined paricalcitol
groups, with a between-group diff erence versus placebo of –15% (95% CI –28 to 1; p=0·071); –14% (from 63 to
54 mg/mmol; –24 to –1) in the 1 μg paricalcitol group, with a between-group diff erence versus placebo of –11%
(95% CI –27 to 8; p=0·23); and –20% (from 61 to 49 mg/mmol; –30 to –8) in the 2 μg paricalcitol group, with a
between-group diff erence versus placebo of –18% (95% CI –32 to 0; p=0·053). Patients on 2 μg paricalcitol showed an
early, sustained reduction in UACR, ranging from –18% to –28% (p=0·014 vs placebo). Incidence of hypercalcaemia,
adverse events, and serious adverse events was similar between groups receiving paricalcitol versus placebo.
Interpretation Addition of 2 μg/day paricalcitol to RAAS inhibition safely lowers residual albuminuria in patients
with diabetic nephropathy, and could be a novel approach to lower residual renal risk in diabetes.
Funding Abbott.
Introduction
Diabetes mellitus represents a worldwide epidemic, and
increases macrovascular and microvascular risk. Drugs
that intervene in the renin–angiotensin–aldosterone
system (RAAS), such as angiotensin-converting enzyme
(ACE) inhibitors or angiotensin receptor blockers
(ARBs), can retard both cardiovascular and renal
morbidity and mortality. This eff ect is partly due to the
lowering of albuminuria associated with RAAS
intervention. However, residual cardiovascular and renal
risk can be extremely high in patients receiving
treatment.1,2 Moreover, residual renal risk is positively
associated with residual albuminuria.3 To address
this residual risk, therapies are needed to further
reduce risk factors, including high blood pressure
and albuminuria.
Strategies to eff ectively inhibit RAAS and interventions
in new pathways have had little success in hard outcome
studies. The combination of an ACE inhibitor and ARB, or
increased doses of each of these drug types have not
reduced cardiovascular or renal risk.4,5 Despite reduction in
albuminuria, endothelin antagonists have an unacceptable
side-eff ect profi le, which led to termination of a trial
investigating hard renal outcomes (composite of doubling
of serum creatinine, end-stage renal disease, or death) with
avosentan.6 Trials with sulodexide, a heparinoid drug, were
stopped because of absence of eff ect.7 New approaches to
stop progressive renal failure are needed.
The natural activator of the vitamin D receptor,
calcitriol, is produced by the kidney, but plasma
concentrations decline as estimated glomerular fi ltration
rate (eGFR) reduces.8 In a multivariable analysis of
patients with chronic kidney disease, lower calcitriol
concentrations strongly correlated with higher risk of
diabetes, higher urinary albumin-to-creatinine ratio
(UACR), and lower eGFR.8 In preclinical models, a

Page 2

Articles
1544
www.thelancet.com Vol 376 November 6, 2010
selective activator of the vitamin D receptor, paricalcitol,
reduced albuminuria and slowed the progression of
kidney injury9–11 with little eff ect on mineral metabolites.
Knockout of the vitamin D receptor in diabetic
mice was associated with severe albuminuria and
glomerulosclerosis from increased thickening of the
glomerular basement membrane
eff acement.12 Moreover, in models of diabetic nephropathy,
combined treatment with paricalcitol and an ARB blocked
the development of albuminuria, maintained the structure
of the glomerular fi ltration barrier, and reduced
glomerulosclerosis in association with reduced renal
expression of renin.13 In a randomised trial investigating
the effi cacy of paricalcitol for lowering of parathyroid
hormone concentrations in patients with stage 3 and 4
chronic kidney disease, fi ndings of a post-hoc analysis
showed that the paricalcitol group had a higher proportion
of patients with signifi cant reductions in semi-quantitative
(dipstick) proteinuria than did the placebo group,14 with
and without RAAS blockade.
We undertook a randomised trial to prospectively test
the eff ectiveness of paricalcitol for the reduction of
residual albuminuria in patients with type 2 diabetic
nephropathy who were receiving stable treatment with
an ACE inhibitor or ARB. Additionally, we studied the
eff ects of dose and time on albuminuria, eGFR, and
blood pressure.
and podocyte
Methods
Patients
Between February, 2007, and October, 2008, we enrolled
patients from 33 hospitals and 27 clinics in Germany,
Greece, Italy, the Netherlands, Poland, Portugal, Spain,
Taiwan, and the USA (including Puerto Rico). Patients
were eligible if they were older than 20 years, had type 2
diabetes and nephropathy, which was defi ned by a UACR
at fi rst morning void of 11–339 mg/mmol and eGFR of
15–90 mL/min per 1·73 m², and had been receiving
stable doses of ACE inhibitors or ARBs for 3 months or
more. Additional inclusion and exclusion criteria,
including serum para thyroid hormone concentration of
35–500 ng/L and serum calcium concentration of less
than 2·45 mmol/L, have been reported previously.15 The
study protocol was approved by an independent ethics
committee and local and central review boards, and all
patients provided written informed consent.
Randomisation and masking
The randomisation sequence was computer generated
by the study sponsor, and was stratifi ed by baseline
eGFR (≤40 vs >40 mL/min per 1·73 m²) and UACR
(≤113 vs >113 mg/mmol). Randomisation was done
centrally with an interactive voice response system
(IVRS) to conceal allocation. Within each stratum,
patients were assigned (1:1:1) to receive 1 μg paricalcitol
once daily, 2 μg paricalcitol once daily, or matching
placebo. Study drugs and packaging were identical in
appearance; the drugs were packaged by the study
sponsor, and distributed to sites through the IVRS as
masked study drug kits. The study investigator, patients,
and study sponsor’s personnel who participated in data
collection and analysis were masked to treatment
assignment for the duration of the study. Treatment
assignment was not disclosed until all database issues
had been resolved and the study database was locked.
Procedures
Patients received their assigned treatment for 24 weeks
(treatment phase) together with their usual care for
diabetes and cardiovascular protection. In accordance
with present guidelines, if a patient’s blood pressure
exceeded 130/80 mm Hg, they could be given an
antihypertensive drug at the physician’s discretion to
obtain acceptable blood pressure control.16 Doses of
ACE inhibitors or ARBs could not be adjusted after
randomisation.
Patients were examined at baseline, every 4 weeks
during the treatment phase, and at 30 and 60 days after
treatment completion (withdrawal phase). Blood pressure
and pulse, adverse events, concomitant drug treatments,
adherence to drug regimens, blood chemistry, and urine
specimens gathered at fi rst morning void on 3 sequential
days were assessed at every visit. Blood samples for
plasma paricalcitol concentrations were obtained at 4, 8,
12, 16, and 20 weeks of treatment; the limit of quantitation
904 patients screened
281 enrolled and randomly
assigned
93 allocated to receive placebo
73 completed the study
88 included primary analysis†
20 discontinued*
2 adverse events
7 withdrew consent
4 lost to follow-up
1 needed dialysis
7 other reasons
15 discontinued*
4 adverse events
3 withdrew consent
1 lost to follow-up
2 needed dialysis
9 other reasons
26 discontinued*
11 adverse events
5 withdrew consent
2 lost to follow-up
4 needed dialysis
13 other reasons
93 allocated to receive 1 μg/day
paricalcitol
78 completed the study
92 included primary analysis†
95 allocated to receive 2 μg/day
paricalcitol
69 completed the study
92 included primary analysis†
623 excluded
571 laboratory values
13 withdrew consent
9 took excluded drug
14 screening period longer than 3 weeks
4 unacceptable medical history
1 investigator’s judgement
11 other reasons
Figure 1: Trial profi le
*Some patients discontinued for more than one reason. †Patients allocated to treatment were excluded from the
analysis if they did not have data for urinary albumin-to-creatinine ratio at baseline and for at least one timepoint
during treatment because of early discontinuation from the study.
See Online for webappendix

Page 3

Articles
www.thelancet.com Vol 376 November 6, 2010 1545
of the assay was 0·01 ng/mL. Additionally, 24 h urine
samples were obtained at baseline, treatment completion,
and 60 days after treatment completion.
The primary effi cacy measure was the percentage
change in geometric mean UACR from baseline to the
last measurement during treatment. Secondary effi cacy
measures recorded from baseline to the last measurement
during treatment included the percentage change in
geometric mean 24 h urinary albumin, and the proportion
of patients achieving at least a 15% reduction in geometric
mean UACR. An additional effi cacy measure recorded
from baseline to the last measurement during treatment
was mean change in geometric mean eGFR. We also
assessed the change in primary and secondary measures
during the 60 days after treatment completion. Further
details of the study procedures are described in
webappendix p 1.
Statistical analysis
We calculated that a total sample size of 258 patients was
needed for at least 82% power to detect an absolute
diff erence in log-transformed UACR of –0·034 mg/mmol
(SD 0·088) from baseline to last measurement during
treatment between the combined paricalcitol group and
placebo at a two-sided signifi cance level of 0·05. The
intention-to-treat dataset included all randomised patients
who had at least one dose of study drug, and was used for
all effi cacy and safety analyses. Patients with missing data
for assessments of change in effi cacy and safety measures
between two timepoints were excluded from analyses
because the change value could not be calculated.
In the protocol-specifi ed primary effi cacy analysis,
ANCOVA was used to compare change in log-transformed
UACR (log UACR) for the combined paricalcitol groups
versus the placebo group, with treatment group as a factor
and baseline UACR as a covariate. Between-group
diff erences in log UACR were reverse transformed and
results were presented as the percentage change in
geometric mean. The secondary effi cacy analysis for
change from baseline to every log UACR after baseline was
a mixed-eff ect, maximum likelihood, repeated measures
analysis. The same ANCOVA and repeated measures
analyses were used to analyse the change from baseline in
other continuous effi cacy variables. A log transformation
was also applied to 24 h albuminuria for the analysis. For
response rate, treatment group diff erences were assessed
with Fisher’s exact test. Diff erences between treatment
groups in the incidence of adverse events were assessed
with Fisher’s exact test. Correlation between change in
UACR and change in other variables was calculated by use
of Spearman’s rank-order method. For safety analyses,
comparisons between treatment groups were done with a
repeated measures mixed model, ANOVA for continuous
variables, or Fisher’s exact test for binary categorical
variables. All analyses were done with SAS (version 9.1.3)
under the Unix operating system. Further details of the
statistical analysis are described in webappendix p 1.
Placebo
(n=93)
1 μg paricalcitol
(n=93)
2 μg paricalcitol
(n=95)
Demographic characteristics
Age (years)
Male sex
Ethnic origin
White
Black
Asian
Other
Current smoker
Body-mass index (kg/m²)
Clinical characteristics
Known duration of diabetes (years)
Blood pressure (mm Hg)
Systolic
Diastolic
Medical history
Hypertension
Hyperlipidaemia
Coronary artery disease
Peripheral vascular disease (arterial)
Congestive heart failure
Urinary albumin-to-creatinine ratio (mg/mmol)
Mean (SD)
Median (IQR)
24 h urinary albumin (mg)
Mean (SD)
Median (IQR)
24 h sodium excretion (mmol)
Mean (SD)
Median (IQR)
Serum albumin (g/L)
Serum creatinine (μmol/L)
eGFR (mL/min per 1·73 m²)
Mean (SD)
Median (IQR)
Haemoglobin (g/L)
HbA1c (mmol/L)
Triglycerides (mmol/L)
Cholesterol (mmol/L)
Total
LDL
HDL
Serum potassium (mmol/L)
Serum calcium (mmol/L)
Serum phosphorous (mmol/L)
Serum intact parathyroid hormone (ng/L)
Mean (SD)
Median (IQR)
1,25-dihydroxy vitamin D (pmol/L)
25-hydroxy vitamin D (nmol/L)*
≥37·44
64 (11)
60 (65%)
64 (10)
66 (71%)
65 (10)
69 (73%)
72 (77%)
11 (12%)
10 (11%)
0
10 (11%)
32 (7)
63 (68%)
15 (16%)
14 (15%)
1 (1%)
6 (6%)
31 (6)
66 (69%)
14 (15%)
15 (16%)
0
14 (15%)
32 (7)
17 (9) 17 (10)16 (9)
142 (17)
73 (12)
142 (18)
73 (12)
141 (16)
73 (9)
92 (99%)
75 (81%)
31 (33%)
18 (19%)
13 (14%)
93 (100%)
69 (74%)
30 (32%)
19 (20%)
18 (19%)
94 (99%)
76 (80%)
28 (29%)
18 (19%)
17 (18%)
94 (81)
73 (30–128)
101 (94)
71 (33–139)
92 (86)
68 (28–133)
1033 (970)
705 (238–1534)
1170 (1373)
597 (314–1724)
1140 (1191)
819 (290–1447)
165 (79)
145 (109–202)
40 (3)
180 (79)
159 (82)
154 (99–200)
40 (4)
172 (56)
161 (73)
146 (105–208)
40 (4)
170 (63)
39 (17)
38 (25–50)
120 (17)
7·6 (1·4)
1·7 (1·0)
40 (15)
37 (27–52)
118 (17)
7·3 (1·3)
1·8 (1·2)
42 (18)
37 (28–55)
121 (16)
7·6 (1·3)
2·2 (1·7)
4·2 (1·2)
2·3 (1·0)
1·1 (0·3)
4·7 (0·7)
2·3 (0·1)
1·2 (0·2)
4·2 (1·1)
2·4 (1·0)
1·1 ((0·3)
4·9 (0·6)
2·3 (0·1)
1·3 (0·2)
4·5 (1·3)
2·5 (1·0)
1·1 (0·4)
4·8 (0·6)
2·3 (0·1)
1·2 (0·2)
105 (91)
66 (48–120)
85 (44)
42 (21)
52 (58%)
97 (77)
75 (46–114)
85 (40)
40 (20)
43 (46%)
91 (65)
70 (49–110)
87 (38)
42 (21)
53 (57%)
(Continues on next page)

Page 4

Articles
1546
www.thelancet.com Vol 376 November 6, 2010
This trial is registered with ClinicalTrials.gov, number
NCT00421733.
Role of the funding source
The study was overseen by a steering committee,
including non-voting members from the sponsor. The
steering committee oversaw the study design, conduct of
the trial, and management and analysis of all data. The
sponsor participated in the study design, collection and
analysis of data, and writing of the report. All authors
had access to study results and the lead author vouches
for the accuracy and completeness of the data reported.
The lead author and the steering committee had the fi nal
decision to submit for publication.
Results
904 patients were screened for eligibility, of whom 281
(31%) were enrolled (fi gure 1). Baseline demographic,
clinical, and biochemical characteristics and concomitant
treatments were balanced between the placebo and
paricalcitol groups (table 1). Overall, 78 patients (28%) had
microalbuminuria, 203 (72%) had macroalbuminuria, and
33 (12%) had eGFR of more than 60 mL/min per 1·73m².
The numbers of patients included in assessments of
change in UACR varied dependent on the timepoints
between which changes were measured. For example,
assessment of UACR from baseline to last measurement
during treatment included 88, 92, and 92 patients,
respectively, in the placebo, 1 μg paricalcitol, and
2 μg paricalcitol groups, whereas assessments from
baseline to week 4 included 87, 85, and 91 patients,
respect ively. As a result, data for the same timepoint vary
between assessments.
In the primary effi cacy analysis, change in geometric
mean UACR from baseline to the last measurement
during treatment was higher in the combined paricalcitol
groups than in the placebo group (table 2), with a
between-group diff erence of –15% (95% CI –28 to 1;
p=0·071; fi gure 2). Change in UACR seemed to have a
dose-response relation with paricalcitol (table 2): in the
1 μg group, the between-group diff erence versus placebo
was –11% (95% CI –27 to 8; p=0·23), and in 2 μg group,
the between-group diff erence versus placebo was –18%
(95% CI –32 to 0; p=0·053).
In the secondary effi cacy analysis, reduction in
geometric mean 24 h rate of albumin excretion was
slightly higher in the 1 μg paricalcitol group than in the
placebo group, with a between-group diff erence of –2%
(95% CI –23 to 25; p=0·86), but was much higher in the 2
μg paricalcitol group than in the placebo group, with a
between-group diff erence of –28% (95% CI –43 to –8;
p=0·009; table 2). The proportion of patients achieving a
change in UACR of at least –15% between baseline and
the last measurement during treatment was higher in the
2 μg paricalcitol group (55% [51/92]) than in the placebo
group (40% [35/88]; p=0·038), but the diff erence between
the 1 μg paricalcitol group (52% [48/92]) and the placebo
group was not signifi cant (p=0·10). With inclusion of all
available data collected across visits, UACR had reduced
substantially by week 4 in the 2 μg paricalcitol group
(–18%, from 60 mg/mmol at baseline to 49 mg/mmol;
95% CI –10 to 9), and this reduction was sustained during
the entire treatment phase, peaking at –28% (reduced to
43 mg/mmol; –37 to –18) at week 12 (fi gure 3A). Across
the treatment phase, reduction in UACR was larger for
2 μg paricalcitol than for placebo (p=0·014; fi gure 3A).
eGFR had also reduced substantially by week 4 in the
2 μg paricalcitol group (p=0·0548 vs placebo), and
Placebo
(n=93)
1 μg paricalcitol
(n=93)
2 μg paricalcitol
(n=95)
(Continued from previous page)
Serum biomarkers
Aldosterone (pmol/L)
Mean (SD)
Median (IQR)
Plasma renin activity (nmol/L per h)
Mean (SD)
Median (IQR)
Brain natriuretic peptide (ng/L)
Mean (SD)
Median (IQR)
Fibrinogen (μmol/L)
Interleukin 6 (ng/L)
Tumour necrosis factor α (ng/L)
Glucose-lowering treatments
Insulin or insulin analogues
Other glucose-lowering drugs
Lipid-lowering treatments
Antihypertensive drugs
ARB
ACE inhibitor
Combination of ARB and ACE inhibitor
Maximum dose of ARB or ACE inhibitor
Combination of ACE inhibitor and other
RAAS inhibitor
Combination of ARB, ACE, and other
RAAS inhibitor
Other RAAS inhibitor
Not on ARB, ACE, or other RAAS inhibitor
Diuretic
Other antihypertensives, calcium channel
blockers, and β blockers
Erythropoetin
Calcium supplements
Aspirin
Other antiplatelet drugs
185 (181)
128 (58–230)
173 (198)
100 (54–211)
171 (172)
125 (61–219)
4·8 (6·1)
2·9 (1·0–5·8)
5·5 (6·8)
2·9 (1·0–5·8)
6·6 (10·4)
3·1 (1·2–5·8)
106 (150)
38 (25–50)
13·4 (3·5)
5·2 (3·2)
5·6 (8·3)
116 (353)
37 (27–52)
14·1 (3·8)
6·4 (14·3)
5·4 (6·8)
88 (104)
37 (28–55)
13·6 (3·4)
5·9 (7·2)
14·5 (79·6)
57 (61%)
58 (62%)
71 (76%)
61 (66%)
49 (53%)
68 (73%)
61 (64%)
64 (67%)
71 (75%)
47 (51%)
24 (26%)
20 (22%)
38 (41%)
2 (2%)
41 (44%)
37 (40%)
14 (15%)
40 (43%)
0
42 (44%)
34 (36%)
14 (15%)
41 (43%)
1 (1%)
0 1 (1%)1 (1%)
0
0
0
0
1 (1%)
2 (2%)
51 (54%)
82 (86%)
53 (57%)
74 (80%)
52 (56%)
74 (80%)
9 (10%)
6 (6%)
42 (45%)
13 (14%)
13 (14%)
5 (5%)
45 (48%)
14 (15%)
9 (9%)
7 (7%)
46 (48%)
15 (16%)
Data are mean (SD) or number (%), unless otherwise indicated. eGFR=estimated glomerular fi ltration rate.
HbA1c=glycosylated haemoglobin. ARB=angiotensin receptor blocker. ACE=angiotensin-converting enzyme.
RAAS=renin–angiotensin–aldosterone system. *Data are missing for three patients in the placebo group and two
patients in the 2 μg paricalcitol group because the centre did not submit the specimen.
Table 1: Characteristics of patients at baseline

Page 5

Articles
www.thelancet.com Vol 376 November 6, 2010 1547
remained stable throughout the treatment phase (range
–3 to –5 mL/min per 1·73 m²; p=0·001 vs placebo;
fi gure 3B), whereas early reduction in systolic blood
pressure was followed by a pattern of fl uctuations across
the treatment phase (range –3 to –9 mm Hg; p=0·033 vs
placebo; fi gure 3C). After controlling for the treatment
eff ect on systolic blood pressure, regression analysis on
log-transformed UACR showed that 84% of eff ect of 2 ug
paricalcitol was direct (p=0·064), and the indirect eff ect
through change in systolic blood pressure accounted for
only 16% (p=0·19). UACR, eGFR, systolic blood pressure,
and 24 h albumin excretion returned towards baseline
values at study visits done at 30 and 60 days after
treatment completion. From the last measurement
during treatment to 60 days after treatment completion,
UACR increased in the 1 μg paricalcitol group, with a
between-group diff erence versus placebo of 43% (95% CI
13 to 82; p=0·001); and in the 2 μg paricalcitol group,
with a between-group diff erence versus placebo of 22%
(95% CI –5 to 55; p=0·076; table 2).
Intact parathyroid hormone increased from 106·2 ng/L
at baseline to 124·5 ng/L at last measurement during
treatment in the placebo group, decreased from 97·5 ng/L
to 70·8 ng/L in the 1 μg paricalcitol group (p<0·0001 vs
placebo), and from 90·7 ng/L to 40·0 ng/L in the 2 μg
paricalcitol group (p<0·0001 vs placebo). No signifi cant
changes were recorded between the paricalcitol groups
and placebo in the measurements of serum C-reactive
protein, fi brinogen, interleukin 6, plasma renin activity,
aldosterone, or tumor necrosis factor α.
Mean plasma concentration of paricalcitol, obtained
from morning fasting plasma samples, was 0·02 ng/mL
(SD 0·024) in the 1 μg paricalcitol group, and 0·03 ng/mL
(SD 0·023) in the 2 μg paricalcitol group; the mean
concentration in the placebo group was below the limit of
quantitation for the assay. We identifi ed an inverse
relation between paricalcitol concentration and UACR
(p<0·0001). In a post-hoc analysis of the eff ects of
–25
–30
–35
PlaceboCombined
paricalcitol
p=0·071
1 μg paricalcitol
p=0·23
2 μg paricalcitol
p=0·053
–20
–15
–10
–5
Change in UACR (%)
0
5
10
15
Figure 2: Change in urinary albumin-to-creatinine ratio from baseline to the
last measurement during treatment
Error bars represent 95% CIs. p values are for the comparison of paricalcitol
versus placebo. UACR=urinary albumin-to-creatinine ratio.
PlaceboCombined
paricalcitol
1 μg paricalcitol2 μg paricalcitol
Geometric mean urinary albumin-to-creatinine ratio
Baseline to last measurement during treatment
Patients
Baseline (mg/mmol)
Last measurement during treatment (mg/mmol)
Percentage change (95% CI)
Last measurement during treatment to 60 days after treatment completion
Patients
Last measurement during treatment (mg/mmol)
60 days after treatment completion (mg/mmol)
Percentage change (95% CI)
Mean 24 h urinary albumin
Baseline to last measurement during treatment
Patients
Baseline (mg)
Last measurement during treatment (mg)
Percentage change (95% CI)
Last measurement during treatment to 60 days after treatment completion
Patients
Last measurement during treatment (mg)
60 days after treatment completion (mg)
Percentage change (95% CI)
88
61
60
–3% (–16 to 13)
184
62
51
–16% (–24 to –9)
92
63
54
–14% (–24 to –1)
92
61
49
–20% (–30 to –8)
72
60
55
–7% (–20 to 8)
139
51
63
23% (11 to 36)
71
57
75
34% (15 to 55)
68
45
52
13% (–3 to 32)
78 146
662
507
–23% (–32 to –13)
74 72
717
463
–34% (–45 to –21)
609
564
–9% (–23 to 8)
613
554
–10% (–25 to 6)
71 143
486
614
25% (12 to 39)
73 70
623
599
–1% (–16 to 16)
531
694
31% (12 to 54)
444
540
19% (1 to 39)
Table 2: Effi cacy analyses