Effects of Intensive Blood Pressure Lowering on the Progression of Chronic Kidney Disease: a Systematic Review and Meta-Analysis

CMAJ Research

Effects of intensive blood pressure lowering on the progression of chronic kidney disease: a systematic review and meta-analysis

Jicheng Lv MD PhD, Parya Ehteshami BSc, Mark J. Sarnak MD, Hocine Tighiouart MS, Min Jun PhD, Toshiharu Ninomiya MD PhD, Celine Foote MD, Anthony Rodgers MD PhD, Hong Zhang MD PhD, Haiyan Wang MD, Giovanni F.M. Strippoli MD PhD, Vlado Perkovic MD PhD

See related commentary by Hildebrand and Garg on page 941 and at www.cmaj.ca/lookup/doi/10.1503/cmaj.130168 Abstract Competing interests: Background: Recent guidelines suggest lower - nine level and 50% decline in glomerular filtra - Jicheng Lv has received ing the target blood pressure for patients with tion rate, or end-stage kidney disease). Com - grant support from Pfizer for hypertension research. chronic kidney disease, although the strength of pared with standard regimens, a more intensive Anthony Rodgers has evidence for this suggestion has been uncertain. blood pressure–lowering strategy reduced the received an unrestricted We sought to assess the renal and cardiovascu - risk of the composite outcome (hazard ratio [HR] grant from Dr. Reddy’s lar effects of intensive blood pressure lowering 0.82, 95% confidence interval [CI] 0.68–0.98) and Laboratories for a trial that in people with chronic kidney disease. end-stage kidney disease (HR 0.79, 95% CI 0.67– includes blood pressure– 0.93). Subgroup analysis showed effect modifica - lowering agents. Vlado Methods: We performed a systematic review Perkovic has received tion by baseline proteinuria ( p = 0.006) and honoraria from Servier for and meta-analysis of all relevant reports pub - markers of trial quality. Intensive blood pressure scientific presentations lished between 1950 and July 2011 identified in lowering reduced the risk of kidney failure (HR relating to blood pressure. a search of MEDLINE, Embase and the Cochrane 0.73, 95% CI 0.62–0.86), but not in patients with - No other competing Library. We included randomized trials that out proteinuria at baseline (HR 1.12, 95% CI interests were declared. assigned patients with chronic kidney disease to 0.67–1.87). There was no clear effect on the risk This article has been peer different target blood pressure levels and re - of cardiovascular events or death. reviewed. ported kidney failure or cardiovascular events. Two reviewers independently identified rele - Interpretation: Intensive blood pressure lower - Correspondence to: Vlado Perkovic, vperkovic vant articles and extracted data. ing appears to provide protection against kid - ney failure events in patients with chronic kid - @george.org.au Results: We identified 11 trials providing infor - ney disease, particularly among those with CMAJ 2013. DOI:10.1503 mation on 9287 patients with chronic kidney dis - proteinuria. More data are required to deter - /cmaj.121468 ease and 1264 kidney failure events (defined as mine the effects of such a strategy among pa - either a composite of doubling of serum creati - tients without proteinuria.

hronic kidney disease is a major public (the Modification of Diet in Renal Disease health problem worldwide, affecting [MDRD] study) that focused on kidney protec - C 10%–15% of the adult population. 1 tion. 11 Subsequent trials of different targets in Blood pressure–lowering agents are the main - people with chronic kidney disease have stay of management strategies aiming to slow yielded inconsistent results, 12,13,14 leading to criti - the progression of chronic kidney disease, as cism by the recent Canadian Hypertension Edu - well as a core aspect of strategies aiming to cation Program guideline (which suggested a reduce cardiovascular risk. 2–4 Observational less aggressive target) of other guidelines, with studies have shown a log-linear increase in the suggestions that their blood pressure recom - risk of kidney failure with high blood pressure mendations went beyond the available evi - levels across the observed range, 5–7 suggesting dence. This criticism has been supported by a that further lowering blood pressure could recent systematic review (no meta-analysis was reduce the risk of kidney failure at most blood performed) that focused on 3 trials and reported pressure levels. Current guidelines recommend inconclusive results overall but raised the possi - a blood pressure target below 130/80 mm Hg bility that proteinuria was an effect modifier. 15 for patients with chronic kidney disease, 8–10 but The final result has been clinician uncertainty this recommendation is mostly based on obser - about optimal blood pressure levels in patients vational studies and a single randomized trial with chronic kidney disease.

We sought to synthesize the results of all blood pressure–lowering agents used, length of available trials that evaluated the effects of dif - follow-up, change in systolic and diastolic blood ferent blood pressure targets in people with pressure during the trial, outcome events and chronic kidney disease and to better define the adverse events. We assessed the risk of bias balance of risks and benefits associated with dif - using the approach recommended by the ferent intensities of blood pressure lowering in Cochrane Collaboration (assessing allocation this population. concealment; blinding of patients, investigators and outcome assessors; completeness; intention- Methods to-treat analysis). For 2 studies, we requested additional information from the original authors Literature search by written correspondence, which we included in We performed a systematic review using the our analysis. 12,14 Any differences in data extrac - approach recommended in the PRISMA (Pre - tion were resolved by discussion among the ferred Reporting Items for Systematic Reviews authors and, where required, adjudication by a and Meta-Analyses) statement. 16 We identified third reviewer (VP). relevant studies by searching MEDLINE via Ovid (from inception through July 2011), Outcomes Embase (from inception through July 2011) and Our primary outcome was a composite of 50% the Cochrane Library database (Cochrane Central decline in glomerular filtration rate and doubling Register of Controlled Trials, no date restriction) of the serum creatinine level, or end-stage kidney for relevant text and medical subject headings disease. We also collected data on major cardio - that included all spellings of “antihypertensive vascular events (i.e., fatal and nonfatal myocar - agents,” “target blood pressure,” “intensive blood dial infarction, fatal and nonfatal stroke) and all- pressure treatment,” “strict blood pressure treat - cause mortality, where available. ment” and “tight blood pressure control” (Appen - dix 1, available at www.cmaj .ca /lookup /suppl /doi Data synthesis and analysis :10.1503/cmaj.121468/-/DC1). We limited our We calculated individual study hazard ratios search to randomized controlled trials (RCTs) (HRs) with 95% confidence intervals (CIs) 13,14,17,18 with at least 6 months’ follow-up, but without age (we extracted relative risks [RRs] when HRs were or language restrictions. We manually scanned not available) for each outcome before pooling. reference lists from identified trials and review For the continuous measurement of blood pres - articles to identify any other relevant studies. In sure, we used the mean difference between addition, we searched the ClinicalTrials .gov web - groups. We obtained summary estimates using a site for randomized trials that were registered as random effects DerSimonian–Laird model. We completed but not yet published. looked for evidence of heterogeneity in treatment The literature search, data extraction and effect between patients with and without baseline quality assessment were done independently by proteinuria by comparing summary results ob - 2 authors using a standardized approach (JL and tained from relevant subgroups (protein–creatinine PE). All completed RCTs that compared more- ratio > 0.22 or daily protein excretion > 300 mg) versus less-intensive blood pressure targets using in a prespecified subanalysis. We estimated the pharmacologic blood pressure–lowering agents percentage of variability across studies attrib - were eligible for inclusion. Data were extracted utable to heterogeneity beyond chance using the I2 either from studies solely involving people with statistic. 19 We made graphic representations of chronic kidney disease or from subsets of other potential publication bias using Begg funnel plots trials where data on patients with chronic kidney of the natural logarithm of the HR versus its stan - disease could be obtained. Chronic kidney dis - dard error and assessed them visually. 20 A 2-sided ease was determined using the definition devel - p value less than 0.05 was considered statistically oped by the Kidney Disease Outcomes Quality significant. Initiative. 8 Results Data extraction and quality assessment We obtained published reports for each trial and Search results and characteristics extracted standard information to a spreadsheet. of included studies The data we sought included baseline patient Our literature search returned 2524 results yield - characteristics (age, sex, systolic and diastolic ing 1650 potentially relevant articles. Of these blood pressure, history of diabetes, history of articles, we reviewed the full text of 67 reports, hypertension and chronic kidney disease), blood from which we identified 12 publications relat - pressure targets in each arm of the study, types of ing to 11 relevant RCTs ( Figure 1 ). These trials

950 CMAJ, August 6, 2013, 185(11) Research involved a total of 9287 participants with chronic Blood pressure targets varied substantially kidney disease (Table 1). 13,14,17,18,21–28 Because kid - between trials. Two trials targeted mean blood ney failure typically develops slowly over time, pressure levels below 92 mm Hg in the intensive we included the long-term post-trial follow-up treatment arm, with a target of 107 mm Hg in the data from the African American Study of Kidney standard treatment arm. 17,26 One trial aimed for Disease and Hypertension (AASK; 8.8–12.2-yr a blood pressure target below 130/80 mm Hg follow-up) 13 and MDRD trials (16.7-yr follow- compared with a diastolic blood pressure of up), although the blood pressure differences 90 mm Hg, 14 1 study targeted below 120/ between the 2 arms were not significant during 80 mm Hg as compared with 135–140/85– the cohort phase in these 2 studies. The reported 90 mm Hg, 28 and 4 studies had diastolic blood trial quality varied substantially. All of the trials pres sure targets below 75–80 mm Hg with com - used open designs, although patients were parators ranging from 80 to 90 mm Hg. 21,23–25 A blinded in 2 trials. Four of the 11 trials did not trial involving pediatric patients targeted a 24- describe allocation concealment, 17,18,21,28 and hour mean blood pressure below the 50th per - another 3 did not describe whether the analysis centile, compared with the 50th to 95th per - was done by intention-to-treat 23–25 (Appendix 2, centiles in the control group. 18 Two trials had more available at www .cmaj .ca /lookup /suppl /doi :10 liberal targets for intensive treatment (< 140– .1503 /cmaj .1214 68 /-/DC1) 150 mm Hg systolic and 85 mm Hg diastolic). 22,29

Articles identifi ed in litera ture search n = 252 4 • MED LINE n = 959 •Embase n = 757 • Cochr ane n = 808

Exc lud ed (dup licates) n = 87 4

Abstract srevie wed n = 1650

Exc lud ed n = 15 85 •Not original inv estigation s (e. g., reviews) (n = 36 8) •Not RC Ts n = 147 Articles identifi ed •Blood press ure lowe ring not ass essed n = 62 from o ther sou rces •No di ff erent blood pressure targets assessed or no n = 2 rele vant ou tcomes n = 93 7 •Trials of hype rtensi on dur ing pregna ncy n = 15 •Stud ies not invo lving human s n = 4 • Mul ti ple pub licat ion s fr om the same tr ial n = 52

Full artic les reviewed n = 67

Exc lud ed n = 56 • No t RC Ts n = 5 • No t original i nv estigation s n = 3 • No re levant ou tcome s n = 32 • Ba seline kidney dis ease status no t ava il able n = 6 •Other pub licati on from the same trial n = 9 • Fo llow-up < 6 mo n = 1

Stu dies includ ed in meta-anal ysis n = 11*

Figure 1: Identification of relevant studies for inclusion in the meta-analysis. RCT = randomized controlled trial. *Included studies were published in 12 articles.

CMAJ, August 6, 2013, 185(11) 951 Research

C t

* . r n 9 5 R R R o e o 2 j n N N N v

2 a

y e

¶ ‡ § ‡ † e . r 2 6 5 2 9 n o u 2 1 1 7 1 l d i n 4 6 1 i a K f

, s

d d d s e c

e e e i u t

0 † t d d d t e i s 1 R l u u u i . b l % l l l r N 5 c c c a e e i x x x t m D E E E c a r

a e l h

% a

c 8 7 7 ,

. . . t x m 8 5 0 1 7 n e e 3 2 4 4 3 s e F i t

, a

, P n

e r a 5 4 2 2 6 g y e 5 5 5 1 5 A m

4 7 8 0 5

9 7 3 4 8 h . 0 D t 3 8 3

o i K 1 n W C

7 4 8 0 5 , l

7 9 3 4 8 . a 0 3 8 3 o

t 1 n o T

p 2 . u

- 2

7 6 0 4 . r

1 . . . w y 6 – 1 5 3 o 1 l 8 l . 8 o F

- s

y e t 1 u

l o n v

m t n R o n i d

m s e o a t h r o

s s y n r h r r v - t l s o t n

p o f t i a g e i : e

e a

d y e n d i c

;

t u h t h a i w l

d r d e d n n r t

r h e i c a t i i p e n r n n o a o a o a c r r t d e n a d

t e r

c o i i a e e h l o a

a a a

p f

o d d

c x v v l d p y o r p e o y p p d i i P i

t l i e e o l h o

o t ; i

p r e ; ; p t t y h

h v v n g g B d r o s

d i i s

i i

n t a p n n

p w a r s h d i t

e t o r e

d e e e e s

t l a w a a g g e d p r k e r r n n

l c c e y

e l i i

i m i e e , l z c n s

o s s e a e e a e e g c i h f e s ; u

r e r t r r r

o a t e e r r u m t f n

l t

w w

a q e e - a o m h

a )

d d e o e o t d d a b d

e t s s

n e

c t α e o - r i o w

g r i m r t t - s y y e n n e

n t r n a d s n l i

b o z c d a a n n s w i d d l

e l i

p n ; a m z i

d e d l l l a e e a t l t p i u u n c

i i i y i i h r o r

e a

i a i t t i e c e m p r r r , s e t t l g s r

e p n s s t r b p P d r t

d n

i e o i r i a a a p p p p e i k a

e e i i i B s e e r o n u 2 2 e g e r d u p p r c i l R c

r r n E i m

a t e d h e e l l s n m m q m a u o × d × e e s o C a l l h t l i y l s m g

l h a a a h e a r 3 w a ( a t p r r A r t r A a A r o a 2 w p a d h c b

p 7 P

u 0 0 – B

n 1 – 9 o 5 7 i

n r

– 8 h 0 < t

a e t g 2

, l 1

e s e i 0 l

c c 0 i i i t g s 5 l l < 1 o

m r

n r y

o o l e a t

r n n e t t a T s s c h n h a a h r t - t n a a e e o

i i e 4 5 5 a c n - M D M 2 i 9 p D 9 a t

e – n

5 o

m n

i H i 6

2 2

a p e

e , t l 9 9 h n e i u

c m t 0 e h i t e t l < < o 0 8 m g

v m n / r

r o 5

r r n n n

e , 0 t g i a e s c

a a P < h 3 t T r - a e e d 1 B i n P e

4 0 i e M < M 2 B p D 8 d u l c

n ,

5 1 0 5 i . . . . / c 0

e i c 5 4 3 6 l s 8

i n / l 9 8 7 7 i e o / / / / l i D 5 t o . 5 7 3 1 e s t d . . . . K 0 s s a u 0 6 8 3 i 3 C a y t

s 5 3 1 2 d 1 s B

h 1 1 1 1 t e i h t

s r r

> 6

d d

n o

)

r 2 e o

– L a n n e e

s a

n e l i p 0 ) d r e d c a a m n

r a u i /

i e p

2 e e i n 2 r t

i l

u m

r g d 3 n d i s a

, p t / e m i a n R i

e n m 7 e i s t r i h r . n

a e m t g i F t

m e n

- i m f , a f / y e e 1 3 a i

p < m i ( ( r

v s G

6 l 3 4 t t

L /

e c

i i

t 7 . w r c r i

8 , r c . –

s s a

2 m L

L L R o d 1 r

n /

e o

y c m 1 y a 1 r 1

n e n o n

F n

d d

L y e c e e r h m p – r / / b v i > h p r o e e

d 0 c o t t

a i G

a i d t t 3 t i r e m i g g i i

m m e s i 0 e r 7

s l n n

d s r

r a h e

i t u f d u c 8 p n e r A e o w n e e 5 m m u n r

u p s C <

i r a l

l – o i n e e

2 2 2 g e l m 4 p

c e n a c : n v c o o t / n s

0 0 5 t i a n i i s i s a h R ) o . . y r r r

t i r

f e L

e a m m m n t < 1 r F e e , l , d t

f 7 7 I e 2 h h h c r e d n m t s t i

o i

/ e / a m 3 – – 3 3 n G D D u

R R p p p t e p r

f n u o s L g e y 7 4 2 7 7 o F F y f o e r K K P e r y . 0 . . . n g . l a r o o n

A h m n c i N n p G 1 3 7 C 1 1 C G 1 B c a a H n c o 1 n 1

)

; t 6 3 t 2 r 1

7 8 l a i l a 1 1 1 t . . a 2 h a

l l

. p

l n t t ( d a a t

a e e e e

t s t

n t u t

l e e e

e n i e 1 h e l y

i r . K l g l t d p g h d o h S i g b a n t p u r ü u

a A l a u t o t t o A

T S C S A ( W R e K W T

952 CMAJ, August 6, 2013, 185(11) Research

, C * t

r R R n o 9 3 6 R e o N N j 2 2 8 n

v N

a 1

, y

. r * * * * ¶ ‡ n o u R R R R 3 l 8 d i n 1 i N N N N a K f

, s

s 9 e c

. i u

8 1 0 0 0 t t e R i s 1 0 0 0 l i b l % N 1 1 1 r a e e i t m D c a r

a e l h

% a

c 5 6 5 ,

. . . t x m 5 4 7 2 5 4 n e e 4 4 4 3 4 6 s e F i t

, a

, . P n

e d r a 1 6 2 8 9 4 e g y e 4 5 6 5 5 7 t A r m o p e

r ,

5 9 9 0 7 h . t 0 7 D 1 9 t 2 3 o

o i 1

6 4 1 1 K n

W C

3 2 =

5 0 8 8 0 0 N ,

7 9 4 1 7 8 , . a 7 1 4 4 4 e o t

t 8

1 4 n

o a

r 1

, o i p t u a

t 0 4 8 0 3 0 r

l . w . . . . . i y f 7 8 3 5 5 2 o

l r l a o

l . F u e r s

s e

e E s m s n

e n o n t o o i i

C o i t e t t l h o

d n

p d

- v p A c i g e i d

d d

n y e a d d i e g d e e h = e a o d

t o z

h z z a n o

r i l i s

, i i a R l c r

d d r r e p

F p

i w a l

m o

m m e e e d

k G u d e o d

o P k y

o o

e d h a o l e a d ,

o n r c t

e e d t v d e B p r d d r r i o

e n i

a v g o o d o i e a a

n i t n n z v

o g e n d l l a r

i e

p n a e a c e a a t l l b d i e e

u s

i e o r e i i e r r c - s s l e t c s g c

a n r r c d - i v h h r s r e β l a u i a o e e e

t t

t a e d b

p p r

r s r

r r r

o - r r e d n o t v a a

s n s d e e n p n e e e r l l o r i o e n t o y

e e

e i n a a o o l n u

o l w s n a w w i r

n l - i t i t l l i d

d . r n n i m , r o o s e a p s s r e d d n p i h r e d

i i i o e e t i t

i y t t a p s

t t t l d g e a p d a i c i m l r r r n d n n

a i o e

a o

a e n a a e e e e n

t i b e e e e r o l = i i n n i p h i i i h l R s r

n h h h d p t a t t d t t i t e l l i o i h i e t t t l l V a y a t n a a m n f d i i i h n n n

C a P e a C A a i t P e i w h I p e e A e w

y , e e r

c s n 9 9

a d 8 8

i u 0 5 m e – –

k s n 6 8 o .

i i

0 0 / r

r 1 e

e d

t < 8 8 0 s e g

s e < s a 4 y l c c c

i 0 a

i i i e g e 1 s t c l l l n o 1 s

r i 0 i s r – / n y i l

o o o - l 0 a 9 t 5 0 d t t t d

e o d i a

9 T s s s n s 3 8 t < n k y – n s a a a

a o

1 1 i i i e e r

5 y c e a c

i - n r r < 8 < D o D D S c n a d o i t

n o

k r e 0 5 0 e

f n

e g h n 8 7 4 o e o i n

c n n m % i H i i 1

g n

l P t

< < 0 p i =

a e

g t e n t B

n < 0 t u s

c c m

0 5 e h a H o i i s 1 e c D i

a c t l l

o - i 8 8 e i g

, t l v m / / r K l b r d r o o

c r

m R n o , 0 0 c t t g C o n i a

t u e

F s s

P 2 5 t , s t m m e . T d s

a a 2

G d D 1 1 e m a B i i n o

r e i 0 y r i r e u m K n o < < D D R d 1 f S r u

i d

s C e 3 R

s u e s 7 l

F e

. h n c

r i n 1 G t

4 l i

n 5 4 , . i

p c i . . r

/ c 6 4 n e

e 5 i . e i c e 8 d 4 4 s l s w 9

i 0 n

/ 9 d

l p 9 o 9 8 a i e o e

/ / / 1 l i s 3 8 t o o e / . / n D . n t 5 l 5 4 e s t r i d i % . . 7 9 3 l s s e c 5 b . a K n 0 u

c 2 6 s i 5 6 a y m n 1 9 t e 5 C e / i a s = 4 3 d 1 1 s

6 L

e d f 1 1 t

P 1 h s e o % i m a t

0 h i % d e , p

0 t 0 0

w 1 r f 5 e

s s

2 g

, s 0 m f o

f < o

s o u u ) e c n

a d 6

y i H f t t r e t o o c s p h l d

i i

z R

– e n i y l l t c p n n s

a l l F o i e

n i e

a n 0 m k a i i m y t t a t t

5 e e a e r G i 2 n g o i t s

r i w s d

s s ≥ 8 i c n

a m n e t o a

H d g r e m m i d s g – i o t i d o y

d i P a e a s n e n 0 i – a e g s s 5 u

p e p e l d i r m t s t t

B t r r p s 6 n 0 m a t c a e e 6 c ,

i p s n

m o

r m

n e ( l t t a - 1 e 8 o ; c

; l l n

d a u m e n n i d t e o o e h n e e d d s m i y

n r l

e c n P - t t c i u v o > t n o o o e o

n ) i o

l b b i k i

5 h l r t a i g i c i o i

n t r B i s a s e a e a s s a a m t g g P 1

s e i p u a

n c h

m i i

r o m s g i e i i c n s s s s / r n n 1 s t B u

c p H H o u d i C t t d d m a

l -

e l d a

a a a L b – r e e

y e r i o y s n

a : c t c o t t t n t

i 0 o 2 2 c y b ) u r d n h y e i s d r r r d d d e m l m m r

n t l

e c s 0

s I r e e v 2 e e e e s e e e o e e e o d

y h h h b c n 1 r 0 m m t t n w l p n n n a t

e p p p

t t t f d p p i i g i i i ; a s e 3 i i i f l u y e P y o y i y f

n f f f y 0 y 9 r t y o y d n i e

e e e e A p w l h y > N d h H B T w 9 T w ( 8 O h s i o n i d t d d d 2 m

o g r r

* e t e e e 5 8 e n t o r r r r 2 2 r t a . . u a

u u u p l l e a l l l l i i i i e = a a h e d p

a a a a v t r (

t t s i E

f f f f

t *

e e a C s s i n 2

y c e y y y 2 * * *

r r e e h A t

w o

e o e e e 3 4 7 e

i i p s i 2 2 2 s e e e y : e y n . . . n n n i i p l s c a l l l D u d d e r d r s b d d d d d

a n y t b i a a a i i i K o u u a h u h t

o u o K r a a i K a t t C K K K t s r t t t c t c

T S N * † ‡ § ¶ S S S U P D S G H e E e S H e

CMAJ, August 6, 2013, 185(11) 953 Research

Effects of intensive blood pressure was small (Appendix 3, available at www.cmaj.ca lowering on renal and cardiovascular /lookup /suppl /doi :10 .1503 /cmaj .121468 /-/DC1). outcomes We did see mild heterogeneity ( I2 = 38.1%, p = 0.1). Further analysis showed that intensive blood Renal outcomes pressure lowering had no effect on kidney failure Seven trials involving 5308 participants recorded in patients who did not have proteinuria (3 trials a total of 1264 kidney failure events. 14,17,18,21,26–28 involving 1218 patients [HR 1.12, 95% CI 0.67– A –7.7-mm Hg difference in systolic blood pres - 1.87]), 13,17,18 but it did reduce the risk of progres - sure and a –4.9-mm Hg difference in diastolic sive kidney failure by 27% (5 trials involving blood pressurem was seen between the 2 treat - 1703 patients [HR 0.73, 95% CI 0.62– ment arms in the studies that reported these dif - 0.86]) 13,14,17,18,27 in people who did have proteinuria ferences. 4,17,21,26–28 Overall, a more intensive regi - at baseline ( Figure 3 ). Subgroup analysis showed men reduced the risk of composite kidney failure heterogeneity between the effect of intensive events by 17% (HR 0.82, 95% CI 0.68–0.98), blood pressure lowering in patients with and and reduced the risk of end-stage kidney disease without proteinuria ( p for heterogeneity = 0.006). alone by 18% (pooled HR for composite out - Excluding the study involving pediatric patients 18 comes 0.79, 95% CI 0.67–0.93) (Figure 2). did not change the overall results, with intensive Formal testing did not identify evidence of blood pressure lowering reducing the risk of kid - publication bias, but the number of included trials ney failure in patients with proteinuria (RR 0.76,

Favou rs Favou rs Treatment, Systolic BP, mm Hg, intensive standard no. of events/patients int ensiv e/ stand ard treatment treatment Study Inte nsiv eSta ndar d Baselin eAchieve d HR (95% CI) Composite outcome Toto et al.21 11/42 7/35 124/ 12 2 133/ 13 8 1.31 (0.57–3.0 2)

Sch rie ret al. 28 5/41 3/34 143/ 14 2 90/1 01 1.38 (0.36–5.3 7)

Ruggenent i et al.14 38/1 69 34/1 69 137.0/136. 4 129.6/133. 7 1.00 (0.61–1.6 4)

Wüh l et al. 18 46/1 89 69/1 96 NA NA 0.65 (0.45–0.9 4)

Appe l et al.13 213/ 54 0 209/ 55 4 152/ 14 9 128/ 14 1 0.95 (0.78–1.1 5)

Hayashi et al.27 5/1230 8/12 69 1717 /17 1.8 135.9/145. 6 0.64 (0.21–1.9 7)

Klahr et al. 17 306/ 43 2 310/ 40 8 130/ 13 1 126.2/133. 8 0.69 (0.59–0.8 3)

Overall (I2 = 38 .1% ) 624/264 3 640/266 5 156.3/156. 4 131.7/141. 5 0.82 (0.68–0 .9 8) ESKD Toto et al. 21 7/42 2/35 124/ 12 2 133/ 13 8 2.92 (0.65– 13.1 5)

Sch rie ret al. 28 5/4 1 3/34 143/ 14 2 90/101 1.38 (0.36–5 .3 7)

Ruggenent i et al. 14 38/1 69 34/1 69 137.0/136. 4 129.6/133. 7 1.00 (0.61–1 .6 4)

Wüh l et al. 18 22/1 89 34/1 96 NA NA 0.67 (0.41–1 .1 0)

Appe l et al. 13 238/ 54 0 256/ 55 4 152/ 14 9 128/ 14 1 0.85 (0.71–1 .0 2)

Hayashi et al. 27 5/12 30 8/12 69 171.7/171. 8 135.9/145. 6 0.64 (0.21–1 .9 7)

Klah r et al. 17 306/ 43 2 310/ 40 8 130/ 13 1 126.2/133. 8 0.69 (0.59–0 .8 3)

Overall (I2 = 21 .6%) 621/264 3 647/266 5 156.3/156. 4 131.7/141. 5 0.79 (0.67–0 .9 3)

0.2 0. 5 1.0 2. 0 5.0 HR (95% CI)

Figure 2: Effects of intensive blood pressure lowering on progressive kidney failure. BP = blood pressure, CI = confidence interval, ESKD = end-stage kidney disease, HR = hazard ratio. Note: Weights are from random effects analysis.

954 CMAJ, August 6, 2013, 185(11) Research

95% CI 0.64–0.89) but not in those without pro - of intensive blood pressure lowering on risk of teinuria (RR 1.03, 95% CI 0.83–1.25) ( p value all-cause death (RR 0.94, 95% CI 0.84–1.05) or for subgroup heterogeneity = 0.03). In addition, cardiovascular death (RR 1.20, 95% CI 0.82– we noticed a larger effect was seen in trials where 1.75) (Appendix 5). allocation concealment and blinding of outcome assessors were not clearly described ( p for het - Potential harms of treatment erogeneity = 0.02 and 0.04, respectively) (Appen - Two trials reported data on severe adverse events dix 4, available at www .cmaj .ca /lookup /suppl /doi (177 events in 723 participants), and no effect on :10 .1503 /cmaj .121468 /-/DC1). the risk of such events was seen with intensive blood pressure lowering (RR 1.04, 95% CI 0.60– Cardiovascular outcomes 1.78). 14,18 Three trials reported the risk of hypoten - Major cardiovascular events were reported in 5 trials sion or associated symptoms. 17,18,26 Individual tri - (472 cardiovascular events in 5308 patients with als reported that more intensive blood pressure chronic kidney disease). 13,14,23 –25 Intensive blood pres - lowering did not increase the risk of hypotension; sure lowering did not reduce the risk of cardiovascu - however, the data were not suitable for meta- lar events in patients with chronic kidney disease, analysis. The risk of acute kidney injury was not but the CIs remained wide (RR 1.09, 95% CI 0.83– reported. Finally, there was no clear difference 1.42). Six trials reported stroke outcomes (197 detected in the rate of stopping treatment between events in 5411 patients), 13,14,22–25 5 trials reported the more- and less-intensive treatment groups in myocardial infarction (138 events in 4317 pa - the 3 trials that reported this outcome. 14,17,18 tients), 14,22–25 and 5 trials reported heart failure (118 events in 5308 patients). 13,14,23–25 We saw no clear Interpretation effect of intensive treatment on any of these vascular outcomes (Appendix 5, available at www .cmaj .ca Most guidelines suggest that blood pressure /lookup /suppl /doi:10 .1503 /cmaj .121468 /-/DC1). should be lowered to below 130/80 mm Hg in all patients with chronic kidney disease. 8,9,30 Our Death meta-analysis of 11 trials involving more than Ten trials involving 6788 participants reported 9000 participants, among whom more than 1200 846 deaths. 13,14,17,18,21–25,28 There was no clear effect kidney failure events were recorded, has shown

Treatment, no. of events/patients Favours intensi ve Favours standard Patient subgroup Inte nsiv eSta ndar d HR (95%CI) treatment treatment Without baseline proteinuria Wüh l et al. 18 NA/3 3 NA/41 1.78 (0.62–5.1 4)

Klahr et al. 17 128/ 21 0 133/ 20 1 0.77 (0.58–1.0 2)

Appe l et al.13 98/3 57 83/376 1.39 (1.04–1.8 6)

Overall (I2 = 78 .2% ) 226/ 60 0 216/ 61 8 1.12 (0.67–1.8 7)

With baseline proteinuria Wüh l et al. 18 NA/6 7 NA/59 0.53 (0.33–0 .8 7)

Ruggenent i et al. 14 38/1 69 34/1 69 1.00 (0.61–1 .6 4)

Klah r et al. 17 178/ 22 1 177/ 20 7 0.71 (0.56–0 .9 1)

Appe l et al. 13 114/ 18 1 126/ 17 6 0.76 (0.58–0 .9 9)

Hayashi et al. 27 2/22 4 4/230 0.51 (0.09–2 .7 8)

Overall (I2 = 0.0%) 332/ 86 2 341/ 84 1 0.73 (0.62–0 .8 6)

0.2 0. 5 1.0 2.0 5.0 HR (95% CI)

Figure 3: Subgroup analysis of the effect of intensive blood pressure lowering on kidney failure in patients with proteinuria compared with those without proteinuria. CI = confidence interval, HR = hazard ratio.

CMAJ, August 6, 2013, 185(11) 955 Research

that intensive blood pressure control with a lower tion to the ESCAPE study with 5 years of follow- target reduced the risk of kidney failure. How - up. 18 It will be critically important for any new tri - ever, the heterogeneity we saw based on markers als assessing interventions aiming to prevent end- of study quality raises the possibility that this stage kidney disease to include participants at benefit is overestimated. high risk for this outcome and to follow them for Clear heterogeneity can be seen in the results of an appropriate period of time to clearly define the trials according to the baseline proteinuria levels of presence or absence of any benefits. participants. Intensive blood pressure lowering A limitation of our study is that most trials of provided clear and consistent protection against different blood pressure–lowering intensities that kidney failure in patients with proteinuria, but not evaluated renal outcomes did not include people in patients without proteinuria. This trend suggests with diabetic kidney disease. Thus, we cannot that aggressive targets may prevent kidney failure answer the question of whether intensive blood in people with proteinuria but not in people with pressure lowering is similarly beneficial in that normal urinary protein excretion, highlighting the population. The Action to Control Cardiovascular importance of assessing urinary protein excretion Risk in Diabetes trial investigated the effects of in people with chronic kidney disease. intensive blood pressure control targeting normal Our results suggest that a 10 mm Hg reduc - systolic pressure (i.e., < 120 mm Hg) in more than tion in systolic blood pressure might result in an 4000 patients with diabetes. 35 Overall, there was no overall reduction of 22% in the risk of kidney effect on end-stage kidney disease in the group failure. This result is consistent with data from receiving intensive treatment, as compared with the observational studies in which lowering systolic conventional systolic blood pressure goal of less blood pressure by 10 mm Hg was associated than 140 mm Hg. However, this trial has not yet with a lower risk of kidney failure. 31–33 reported the effects of intensive blood pressure low - The results from individual trials were incon - ering according to levels of proteinuria, although sistent, even among trials with the same target these data would be of great interest and value. blood pressure, highlighting the value of meta- Our review was also limited by the varied qual - analysis in improving precision and providing an ity of the included trials and the inability to use a opportunity to identify important risk modifiers double-blinded study design for a trial of blood with increased statistical power. 12,13,17,26 Our results pressure targets, raising the risk of differential use suggest clinical heterogeneity in the study popu - of other treatments. In addition, the heterogeneity lation may explain the variable results across we saw according to markers of study quality lim - these trials, because proteinuria was found to be its the strength of the conclusions that can be a clear effect modifier. drawn. Furthermore, the number of included trials The pooled effects of intensive blood pressure is small, which limited subgroup analysis, and the lowering on renal outcomes were consistent various definitions used for blood pressure targets among studies and subgroups, including partici - made meaningful subgroup analyses by different pants with proteinuria, with no evidence of het - target levels impossible. Finally, because adverse erogeneity. These results were consistent with events were inconsistently reported, safety re - those from the individual studies that reported an mains an area of uncertainty. interaction between blood pressure lowering and We did not find any evidence of benefit from proteinuria. 11,13 The mechanisms underlying this intensive blood pressure lowering on cardiovascu - observation are unclear, but patients with pro - lar events or death. Cardiovascular outcomes were teinuria could have high intraglomerular pressure not reported in all of the trials, and the total num - and may therefore be particularly sensitive to bers of events identified were modest, suggesting intensive blood pressure lowering. that this may be an issue of power. Reviews of dif - ferent blood pressure–lowering targets in the gen - Strengths and limitations eral population have suggested that overall cardio - A key strength of our review is its inclusion of vascular benefit is achieved with lower blood long-term follow-up data from a number of trials pressure targets 36 and have refuted concerns with large numbers of end-stage kidney disease derived from observational studies that lower tar - events, thus providing substantial statistical power gets in the commonly observed range may lead to to detect plausible benefits. Kidney function gen - increases in the risk of coronary artery disease. erally declines relatively slowly, highlighting the importance of long-term follow-up to obtain ad - Conclusions equate power to show clinically important benefits. The available data suggest that intensive blood Our meta-analysis includes 2 trials involving pressure lowering prevents end-stage kidney dis - patients at high risk for end-stage kidney disease ease in patients with chronic kidney disease, par - with more than 10 years of follow-up, 17,34 in addi - ticularly in the presence of proteinuria. However,

956 CMAJ, August 6, 2013, 185(11) Research the effects remain to be shown in people with 22. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. UK normal urinary protein excretion. Our data pro - Prospective Diabetes Study Group [published erratum in BMJ vide support for current guidelines that suggest 1999; 318:29]. BMJ 1998;317:703-13. 23. Hansson L, Zanchetti A, Carruthers SG, et al. Effects of inten - lowering blood pressure targets in people with sive blood-pressure lowering and low-dose aspirin in patients proteinuric chronic kidney disease and highlight with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. HOT Study Group. Lancet the potential value of this approach. More data 1998; 351:1755-62. are required to determine the effect of such a 24. Estacio RO, Jeffers BW, Gifford N, et al. Effect of blood pres - strategy in patients who do not have proteinuria. sure control on diabetic microvascular complications in patients with hypertension and type 2 diabetes. Diabetes Care 2000;23 (Suppl 2):B54-64. References 25. Schrier RW, Estacio RO, Esler A, et al. Effects of aggressive blood pressure control in normotensive type 2 diabetic patients on albu - 1. James MT, Hemmelgarn BR, Tonelli M. Early recognition and minuria, retinopathy and strokes. Kidney Int 2002; 61: 1086-97. prevention of chronic kidney disease. Lancet 2010;375:1296-309. 26. Wright JT Jr, Bakris G, Greene T, et al. Effect of blood pressure 2. Heerspink HJ, Ninomiya T, Perkovic V, et al. Effects of a fixed lowering and antihypertensive drug class on progression of combination of perindopril and indapamide in patients with type hypertensive kidney disease: results from the AASK trial. JAMA 2 diabetes and chronic kidney disease. Eur Heart J 2010; 31: 2002; 288:2421-31. 2888-96. 27. Hayashi K, Saruta T, Goto Y, et al. Impact of renal function on 3. Ruilope LM, Salvetti A, Jamerson K, et al. Renal function and cardiovascular events in elderly hypertensive patients treated intensive lowering of blood pressure in hypertensive participants with efonidipine. Hypertens Res 2010;33:1211-20. of the hypertension optimal treatment (HOT) study. J Am Soc 28. Schrier R, McFann K, Johnson A, et al. Cardiac and renal effects Nephrol 2001;12:218-25. of standard versus rigorous blood pressure control in autosomal- 4. Jafar TH, Stark PC, Schmid CH, et al. Progression of chronic dominant polycystic kidney disease: results of a seven-year kidney disease: the role of blood pressure control, proteinuria, prospective randomized study. J Am Soc Nephrol 2002;13:1733-9. and angiotensin-converting enzyme inhibition: a patient-level 29. JATOS Study Group. Principal results of the Japanese trial to meta-analysis. Ann Intern Med 2003;139:244-52. assess optimal systolic blood pressure in elderly hypertensive 5. Klag MJ, Whelton PK, Randall BL, et al. Blood pressure and patients (JATOS). Hypertens Res 2008;31:2115-27. end-stage renal disease in men. N Engl J Med 1996;334:13-8. 30. Mancia G, De Backer G, Dominiczak A, et al. 2007 Guidelines 6. Tozawa M, Iseki K, Iseki C, et al. Blood pressure predicts risk of for the Management of Arterial Hypertension: The Task Force developing end-stage renal disease in men and women. Hyper - for the Management of Arterial Hypertension of the European tension 2003;41:1341-5. Society of Hypertension (ESH) and of the European Society of 7. Perkovic V, Huxley R, Wu Y, et al. The burden of blood pres - Cardiology (ESC). J Hypertens 2007;25:1105-87. sure-related disease: a neglected priority for global health. 31. Lewington S. Blood pressure, cholesterol and premature death: Hypertension 2007;50:991-7. towards the real relationship . Oxford (UK): University of Oxford; 8. K/DOQI clinical practice guidelines for chronic kidney disease: 1999. evaluation, classification, and stratification. Am J Kidney Dis 32. Klag MJ WP, Randall BL, Neaton JD, et al. Blood pressure and 2002; 39(Suppl 1):S1-266. end-stage renal disease in men. N Engl J Med 1996;334:13-8. 9. Chobanian AV, Bakris GL, Black HR, et al. The seventh report 33. O’Seaghdha CM PV, Lam TH, McGinn S, et al.; Asia Pacific of the Joint National Committee on Prevention, Detection, Eval - Cohort Studies Collaboration. Blood pressure is a major risk fac - uation, and Treatment of High Blood Pressure: the JNC 7 report. tor for renal death: an analysis of 560 352 participants from the JAMA 2003;289:2560-72. Asia-Pacific region. Hypertension 2009;54:509-15. 10. Mancia G, Laurent S, Agabiti-Rosei E, et al. Reappraisal of 34. Appel LJ, Anderson CA. Compelling evidence for public health European guidelines on hypertension management: a European action to reduce salt intake. N Engl J Med 2010;362:650-2. Society of Hypertension Task Force document. J Hypertens 35. Gerstein HC, Miller ME, Byington RP, et al. Effects of intensive 2009; 27:2121-58. glucose lowering in type 2 diabetes. N Engl J Med 2008; 358: 11. Peterson JC, Adler S, Burkart JM, et al. Blood pressure control, 2545-59. proteinuria, and the progression of renal disease. The Modification 36. Lv J, Neal B, Ehteshami P, et al. Effects of intensive blood pres - of Diet in Renal Disease study. Ann Intern Med 1995; 123: 754-62. sure lowering on cardiovascular and renal outcomes: a system - 12. Sarnak MJ, Greene T, Wang X, et al. The effect of a lower target atic review and meta-analysis. PLoS Med 2012;9:e1001293. blood pressure on the progression of kidney disease: long-term follow-up of the modification of diet in renal disease study. Ann Affiliations: The George Institute for Global Health (Lv, Intern Med 2005;142:342-51. Ehteshami, Jun, Foote, Rodgers, Perkovic), The University of 13. Appel LJ, Wright JT Jr, Greene T, et al. Intensive blood-pressure control in hypertensive chronic kidney disease. N Engl J Med Sydney, Sydney, Australia; Renal Division, Department of 2010; 363:918-29. medicine (Lv, Zhang, Wang), Peking University First Hospi - 14. Ruggenenti P, Perna A, Loriga G, et al. Blood-pressure control tal, Beijing, China; Tufts Medical Center (Sarnak, for renoprotection in patients with non-diabetic chronic renal Tighiouart), Boston, Mass.; the Department of Medicine and disease (REIN-2): multicentre, randomised controlled trial. Clinical Science Graduate School of Medical Sciences Lancet 2005;365:939-46. (Ninomiya), Kyushu University, Japan; the Centre for Kid - 15. Upadhyay A, Earley A, Haynes SM, et al. Systematic review: ney Research (Strippoli), The Children’s Hospital at West - blood pressure target in chronic kidney disease and proteinuria mead, School of Public Health, The University of Sydney, as an effect modifier. Ann Intern Med 2011;154:541-8. 16. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement Sydney, Australia; the Department of Pharmacology and Epi - for reporting systematic reviews and meta-analyses of studies demiology (Strippoli), Mario Negri Sud Consortium, S Maria that evaluate health care interventions: explanation and elabora - Imbaro, Chieti, Italy; and DIAVERUM Medical-Scientific tion. Ann Intern Med 2009;151:W65-94. Office (Strippoli), Lund, Sweden 17. Klahr S, Levey AS, Beck GJ, et al. The effects of dietary protein restriction and blood-pressure control on the progression of Contributors: Vlado Perkovic, Jicheng Lv and Parya Ehte - chronic renal disease. Modification of Diet in Renal Disease shami conceived the idea for the study. Jicheng Lv, Parya Ehte - Study Group. N Engl J Med 1994;330:877-84. shami, Mark J. Sarnak, Hocine Tighiouart and Vlado Perkovic 18. Wühl E, Trivelli A, Picca S, et al. Strict blood-pressure control collected, analyzed and interpreted the data. Jicheng Lv, Parya and progression of renal failure in children. N Engl J Med 2009; Ehteshami, Vlado Perkovic drafted the manuscript. All of the 361: 1639-50. authors revised the manuscript for important intellectual con - 19. Woodard M. Epidemiology: design and data analysis . 2nd ed. Boca Raton (FL): Chapman & Hall/CRC Press; 2005;687-90. tent and approved the final version submitted for publication. 20. Egger M, Davey Smith G, Schneider M, et al. Bias in meta- Funding: Jicheng Lv was supported by an Amgen Renal analysis detected by a simple, graphical test. BMJ 1997; 315: 629-34. Research Fellowship. Vlado Perkovic was supported by an 21. Toto RD, Mitchell HC, Smith RD, et al. “Strict” blood pressure Australian Heart Foundation Career Development Award. The control and progression of renal disease in hypertensive nephro - funders had no role in study design, data collection and analy - sclerosis. Kidney Int 1995;48:851-9. sis, decision to publish, or preparation of the manuscript.

CMAJ, August 6, 2013, 185(11) 957