Articles. Funding Boehringer-Ingelheim.

Transcription

1 Renal outcomes with telmisartan, ramiril, or both, in eole at high vascular risk (the ONTARGET study): a multicentre, randomised, double-blind, controlled trial Johannes F E Mann, Roland E Schmieder, Matthew McQueen, Leanne Dyal, Helmut Schumacher, Janice Pogue, Xingyu Wang, Aldo Maggioni, Andrzej Budaj, Suhachai Chaithirahan, Kenneth Dickstein, Matyas Keltai, Kaj Metsärinne, Ali Oto, Alexander Parkhomenko, Leooldo S Piegas, Tage L Svendsen, Koon K Teo, Salim Yusuf, on behalf of the ONTARGET investigators Summary Background Angiotensin recetor blockers (ARB) and angiotensin converting enzyme (ACE) inhibitors are known to reduce roteinuria. Their combination might be more effective than either treatment alone, but long-term data for comarative changes in renal function are not available. We investigated the renal effects of ramiril (an ACE inhibitor), telmisartan (an ARB), and their combination in atients aged 55 years or older with established atherosclerotic vascular disease or with diabetes with end-organ damage. Methods The trial ran from 21 to 27. After a 3-week run-in eriod, articiants were randomly assigned to ramiril 1 mg a day (n=8576), telmisartan 8 mg a day (n=8542), or to a combination of both drugs (n=852; median follow-u was 56 months), and renal function and roteinuria were measured. The rimary renal outcome was a comosite of dialysis, doubling of serum creatinine, and death. Analysis was by intention to treat. This study is registered with ClinicalTrials.gov, number NCT Findings 784 atients ermanently discontinued randomised theray during the trial because of hyotensive symtoms (46 on combination theray, 149 on ramiril, and 229 on telmisartan). The number of events for the comosite rimary outcome was similar for telmisartan (n=1147 [13 4%]) and ramiril (115 [13 5%]; hazard ratio [HR] 1, 95% CI ), but was increased with combination theray (1233 [14.5%]; HR 1 9, , = 37). The secondary renal outcome, dialysis or doubling of serum creatinine, was similar with telmisartan (189 [2 21%]) and ramiril (174 [2 3%]; HR 1 9, ) and more frequent with combination theray (212 [2 49%]: HR 1 24, , = 38). Estimated glomerular filtration rate (egfr) declined least with ramiril comared with telmisartan ( 2 82 [SD 17 2] ml/min/1 73 m² vs 4 12 [17 4], < 1) or combination theray ( 6 11 [17 9], < 1). The increase in urinary albumin excretion was less with telmisartan (= 4) or with combination theray (= 1) than with ramiril. Interretation In eole at high vascular risk, telmisartan s effects on major renal outcomes are similar to ramiril. Although combination theray reduces roteinuria to a greater extent than monotheray, overall it worsens major renal outcomes. Funding Boehringer-Ingelheim. Introduction Clinical trials of angiotensin-converting enzyme (ACE) inhibitors and angiotensin recetor blockers (ARB) have shown that these drugs reduce albuminuria, as well as renal risk ie, loss of glomerular filtration rate (GFR) and need for dialysis in those with advanced renal disease. 1 Whether a combination of these classes of drugs further reduces renal risk, and whether the effects of ARB and ACE inhibitors are equivalent, is unknown. A recent meta-analysis indicated that the combination of an ACE inhibitor and an ARB reduces roteinuria to a greater extent than either drug alone, 2 but the total number of atients in each trial was less than 3 on average, the duration of theray rarely exceeded 1 year, and the effects on major renal outcomes, such as GFR changes or occurrence of dialysis, was not reorted. Proteinuria has been suggested not only as a marker but also as a cause of rogressive renal insufficiency. 3 Further reduction of roteinuria by combined ACE inhibitor and ARB theray could theoretically rotect the kidney from chronic kidney failure comared with either agent alone. 3 Proteinuria has also been linked with increased cardiovascular morbidity. 4 Combining ACE inhibitors with an ARB might also lead to more frequent adverse events than monotheray, 5 including acute renal failure and hyerkalaemia. Assessment of the net balance of benefits and risks of ACE inhibitors and ARB alone and in combination requires direct comarisons in large randomised controlled trials. Although most trials of renal disease rogression include articiants with re-existing advanced renal disease, these atients constitute only a small number of atients who are eligible to receive ACE inhibitors and ARB. As art of the resecified analyses of the ONgoing Alone and in combination with Ramiril Lancet 28; 372: See Comment age 511 See Persectives age 527 Grochowski Hosital, Postgraduate Medical School, Poland (A Budaj MD); Cardiac Centre, Chaohya Hosital, Bangkok, Thailand (Prof S Chaithirahan MD); Division of Cardiology, University of Bergen Stavanger Hosital, Norway (K Dickstein MD); Poulation Health Research Institute, McMaster University, Hamilton, Canada (L Dyal MSc); Gottsegen György Hungarian Institute of Cardiology, Semmelweis University, Budaest, Hungary (Prof M Keltai MD); ANMCO Research Centre, Firenze, Italy (A Maggioni MD); Schwabing General Hosital, and KfH Kidney Centre, Ludwig Maximilians University Munchen, Germany (Prof J F E Mann MD); Poulation Health Research Institute, McMaster University, Hamilton, Canada (Prof M McQueen MD); Turku University Central Hosital/ Kantasairaala, Finland (K Metsärinne MD); Cardiology, Hacettee University, Ankara, Turkey (Prof A Oto MD); Institute of Cardiology, Kiev, Ukraine (A Parkhomenko MD); Instituto Dante Pazzanese De Cardiologia, Brazil (L S Piegas MD); Poulation Health Research Institute, McMaster University, Hamilton, Canada (J Pogue PhD); Deartment of Nehrology and Hyertension, Friedrich Alexander University Erlangen, Germany (Prof R E Schmieder MD); Boehringer Ingelheim, Germany (H Schumacher MD); Holbaek City Hosital, Denmark (T L Svendsen MD); Poulation Health Research Institute, McMaster University, Vol 372 August 16,

2 Hamilton, Canada (Prof K K Teo MD); Beijing Hyertension League InstituteFu Wai Hosital, Beijing, China (X Wang PhD); and Poulation Health Research Institute, McMaster University, Hamilton, Canada (Prof S Yusuf MD) Corresondence to: Johannes FE Mann, ONTARGET Office, McMaster University, Hamilton General Hosital, 237 Barton Street East, Hamilton, ONT L8L 2X2 Canada tbn2ab@mail.lrz-muenchen. de 8542 allocated telmisartan 8 mg a day Global Endoint Trial (ONTARGET) study, 6 we examined the effects of telmisartan, (an ARB), ramiril (an ACE inhibitor), and their combination used at full doses, on renal outcomes in a large oulation at high cardiovascular risk. We secifically reort: the frequency of the rimary comosite renal outcome of dialysis, doubling of serum creatinine, and death, and each comonent; and changes in surrogate markers such as estimated GFR (egfr) and roteinuria. Methods Study design As described reviously, 7 we enrolled articiants who were aged 55 years or older with established atherosclerotic vascular disease or with diabetes with endorgan damage atients, single-blind run-in, days randomised to double-blind treatment 8576 allocated ramiril 1 mg a day 3399 excluded 1123 oor comliance 597 withdrew 492 symtoms of hyotension 223 hyerkalaemia 64 elevated s-creatinine 27 death 872 other reasons 852 allocated telmisartan 8 mg a day lus ramiril 1 mg a day Figure 1: Trial rofile All articiants received randomised theray excet 43 lost to follow-u, so 99 8% were followed-u until end of study or until a rimary event. 6 Decrease in egfr from run-in Ramiril and telmisartan Ramiiril Run-in Week 6 Year 2 Time eriod Study end Figure 2: Decrease in estimated glomerular filtration rate (egfr) during the trial, from baseline to study end Exclusion criteria included major renal artery stenosis, uncorrected volume or sodium deletion, a serum creatinine concentration above 265 μmol/l, and uncontrolled hyertension (>16 mm Hg systolic or >1 mm Hg diastolic). Before entering the study, 64 1% of articiants were taking ACE inhibitors or ARB. 7 After a 3-week single-blind run-in eriod, articiants were randomly assigned (double-blind) to receive telmisartan 8 mg a day, ramiril 1 mg a day, or their combination. Particiants were randomised in blocks and stratified by centre with an automated system via telehone. 7 During run-in, all articiants received single-blind study medication for 3 4 weeks in increasing doses of 5 mg ramiril lus 4 mg telmisartan (forced titration). Of articiants at run-in, 64 ( 2%) withdrew because of an elevated serum creatinine (figure 1). Baseline characteristics of the articiants have been ublished reviously articiants were randomised and followed-u after 6 weeks, and then every 6 months thereafter, for a median of 56 months. Serum creatinine was measured before run-in, 6 weeks after randomisation, after 2 years, and at the study end. The urine albumin to creatinine ratio was measured before run-in, at 2 years, and at the enultimate visit, and a value between 3 4 mg/mmol and below 33 9 mg/mmol was defined as microalbuminuria. A value of 33 9 mg/mmol or more (roughly 3 mg/g creatinine) was defined as. Urine albumin was measured centrally by a turbidimetric method (Unicel DxC6 Synchron Systems, Beckman Coulter, Bea, CA, USA). The coefficient of variation at 32 2 mg/l was 4 4% and at 15 5 mg/l was 2 4%. A human serum ool at a concentration of 1 9 mg/l gave a coefficient of variation of 9 2%, and at mg/l gave a coefficient of variation of 2 7%. Creatinine in urine was measured centrally by a modified Jaffe method (Unicel DxC6 Synchron Systems). The coefficient of variation at 7 9 mg/l was 2 9%, and at 23 1 mg/l was 2 8%. A human serum ool at a concentration of 1 8 mg/l gave a coefficient of variation of 2 6%, and at 13 mg/l gave a coefficient of variation of 1 8%. Serum creatinine was measured locally at the study sites. From the serum creatinine, egfr was calculated using the four-variable Modification of Diet in Renal Disease formula. 8 Serum creatinine values below 35 4 μmol/l (n=44 at baseline, 59 at follow-u) or above 16 8 μmol/l (n= at baseline, two at follow-u) were deemed imlausible and excluded from the analysis. Information about dialysis was asked at each visit. After trial conclusion, a questionnaire was sent to all sites having reorted instances of dialysis. This questionnaire required information about duration of dialysis and rimary reasons for any acute dialysis. In three of 162 instances, we did not get information on whether dialysis was acute or chronic; investigators could reort several reasons for acute dialysis. Before the unblinding of the study, we develoed a statistical analysis lan for the renal outcomes. The Vol 372 August 16, 28

3 rimary comosite outcome was the first occurrence of any dialysis, renal translantation, the doubling of serum creatinine, or death; however, no cases of renal translant were reorted in this trial. The secondary renal outcomes included the comosite of any dialysis and doubling of serum creatinine, comonents of the comosite outcomes, changes in egfr and rogression of roteinuria (defined as new develoment of micro albuminuria or ). Investigators reorted cases of renal imairment leading to intermittent or ermanent discontinuation of trial medication. Renal imairment was not defined but was based on reorts from clinical investigators stating a reason for discontinuing trial medication. Acute and chronic dialysis was defined on duration of renal relacement theray of 2 months or less or of more than 2 months. Investigators were asked to rovide the rimary reason for any acute dialysis. Statistical analysis Continuous data are given as mean (SD) and categorical data as actual frequencies and ercentages. The rimary analysis used a time-to-event aroach and included all randomised articiants (intention to treat). Treatment comarisons with regard to time-to-event related data (based on Cox regression of time to occurrence of first event) are shown as hazard ratio (HR) with 95% CI. Analyses on resecified subgrous were done using the Cox regression model, with factors for treatment, subgrou, and interactions. All values were two-sided. Treatment grou comarisons for categorical data were done with the χ² test. For continuous variables comarisons between treatment grous were done with t tests. Urine albumin levels and urine albumin-creatinine ratios were not normally distributed, therefore values were log-transformed before any analysis. For log-transformed data the geometric mean and 95% CI are resented. This study is registered with ClinicalTrials. gov, number NCT Role of the funding source The trial was coordinated and analysed indeendently by the Poulation Health Research Institute (PHRI) at McMaster University and the steering committee had overall resonsibility for interretation of results. The database was transferred to the sonsor at the end of the study. Results At baseline, mean serum creatinine was 93 7 (SD 22 8) μmol/l and mean egfr 73 6 (19 6) ml/min/1 73 m²; egfr was less than 6 ml/min/1 73 m² in 6157 articiants (mean 5 7 [7 7]) and 6 ml/min/1 73 m² or more in articiants (8 9 [15 6]). There were few articiants with an egfr below 3 ml/min/1 73 m² (n=263). Creatinine clearances calculated according to the Cockcroft Gault formula were higher than egfr by 6 8% (data not shown). During follow-u, egfr concentrations Ramiril Ramiril+ telmisartan vs ramiril egfr, baseline 73 7 (19 3) 73 6 (19 9) 73 4 (19 5) egfr change 2 14 (12 9) 2 51 (13 2) 4 1 (13 3) 7 < 1 baseline to 6 weeks egfr change 1 96 (15 1) 3 5 (15 1) 5 12 (15 7) < 1 < 1 baseline to 2 years egfr change 2 82 (17 2) 4 12 (17 4) 6 11 (17 9) < 1 < 1 6 baseline to final egfr change 6 weeks to final 1 17 (17 1) 2 6 (17 1) 2 49 (17 4) 32 < 1 decreased more with telmisartan and with combination theray than with ramiril alone (figure 2, table 1). Similarly, serum creatinine showed greater increases with combination theray than with ramiril (data not shown). The frequency of the comosite rimary renal outcome of dialysis, doubling of serum creatinine, and death was similar with telmisartan (1147, 13 4%) and ramiril (115, 13 5%; HR 1, 95% CI ), but increased with combination theray (1233, 14 5%; HR 1 9, ; = 37). The secondary renal endoint of dialysis or doubling of serum creatinine was similar with telmisartan (189, 2 21%) and ramiril (174, 2 3%; HR 1 9, ) but was more frequent with combination theray (212, 2 49%; HR 1 24, , = 38; figure 3, table 2). There were 162 cases of dialysis, of Ramiril+telmisartan vs ramiril egfr=estimated glomerular filtration rate (ml/min/1 73 m2 [SD]). Number of articiants with measurements= at baseline, at 6 weeks, at 2 years, at study end. Table 1: Estimated glomerular filtration rate at baseline and changes of egfr All dialysis, doubling, death All dialysis and doubling Ramiril n (%) 115 (13 4) 174 (2 3) All dialysis 48 ( 56) All death 114 (11 8) Doubling 14 (1 63) Acute dialysis Chronic dialysis 13 ( 15) 33 ( 39) n (%) 1147 (13 4) 189 (2 21) 51 ( 6) 989 (11 6) 155 (1 81) 2 ( 23) 31 ( 36) Ramiril+ telmisartan n (%) 1233 (14 5) 212 (2 49) 63 ( 74) 165 (12 5) 166 (1 95) 28 ( 33) 34 ( 4) vs ramiril HR 1 ( ) 1 9 ( ) 1 7 ( ) 98 ( 9 1 7) 1 11 ( ) 1 55 ( ) 94 ( ) Ramiril+ telmisartan vs ramiril HR ( ) ( ) ( ) ( ) ( ) ( ) ( ) Dialysis=at least one dialysis. Chronic dialysis=more than 2 months. Acute dialysis=2 months or less. Doubling=doubling of serum creatinine from baseline values. HR=hazard ratio. Reasons for acute dialysis were reorted as severe infection (n=22), volume deletion (n=9), ost-surgery (n=7), drugs (n=5), secific renal diseases (n=5), and other reasons (n=23). In three of 165 originally reorted cases of dialysis, 6 detailed analysis revealed that no dialysis took lace. In three of the 162 cases of dialysis, we got no information on duration of dialysis. Investigators could reort several reasons for acute dialysis. Table 2: Incidence of rimary and secondary renal outcomes and of its comonents Vol 372 August 16,

4 Cumulative incidence rates Number at risk Ramiril and ramiril Cumulative incidence rates Number at risk Ramiril and ramiril Cumulative incidence rates Number at risk Ramiril and ramiril A B C and ramiril Ramiril Years of follow-u T vs R = 968 R+T vs R = Years of follow-u T vs R = 42 R+T vs R = T vs R = 747 R+T vs R = Years of follow-u Figure 3: Kalan-Meier curves for (A) rimary renal outcome (dialysis, doubling of serum creatinine, and death), (B) secondary renal outcome (dialysis and doubling of serum creatinine), and (C) dialysis For statistical details, see table 2. T=telmisartan. R=ramiril which 61 were acute dialyses and which were more frequent for combination theray than for ramiril (table 2). The incidence of chronic dialysis (n=98) was similar between grous; in three cases we got no information on duration of dialysis. Removing acute dialysis from the analysis did not substantially alter the rimary renal outcome; numbers of the secondary renal outcome were reduced to 16 for articiants on ramiril, 172 on telmisartan, and 185 on combination theray (= 141 for ramiril vs combination theray). In all subgrous, telmisartan and ramiril had similar effects on the rimary renal outcome (figure 4). Combination theray had no clear benefit in the highest renal risk grou (overt diabetic nehroathy), in articiants with hyertension and diabetes, or in articiants with an egfr below 6 ml er min, but tended to be harmful in some individuals with low renal risk eg, those without hyertension or diabetes. When the comosite of doubling of serum creatinine or dialysis or the individual comonents of the comosite were analysed searately, combination theray showed no benefit in high renal risk grous but tended to show worse results in low renal grous (data not shown). 784 atients ermanently discontinued randomised theray during the trial because of hyotensive symtoms (46 eole on combination theray, 149 on ramiril, in 229 on telmisartan). Renal abnormalities were reorted in 871 (1 2%) articiants assigned to ramiril, 96 (1 6%) assigned to telmisartan, and 1148 (13 5%) assigned to combination theray (RR 1 33, 95% CI , < 1 for combination theray vs ramiril), and as a reason for ermanently stoing medication in 6 ( 7%) of articiants assigned to ramiril, 68 ( 8%) assigned to telmisartan, and 94 (1 1%) assigned to combination theray (RR 1 58, < 5 for combination theray vs ramiril). Table 3 shows urinary albumin excretion at baseline and its changes during the trial. The geometric mean of urine albumin excretion at baseline ranged between 81 mg/mmol and 83 mg/mmol creatinine and was not different between the randomised grous. Microalbuminuria was resent in 13 1% of all articiants, in 29 7% of those with diabetes, and in 9 2% of those without known diabetes. Macroalbuminuria was seen in 4 % of all articiants, in 12 2% of those with diabetes, and in 1 4% of those without known diabetes. Urine albumin increased at 2 years and at study end to a lesser extent in articiants assigned telmisartan or combination theray than in those assigned ramiril. The geometric mean of the urine albumin to creatinine ratios increased from baseline to last observation by 31% (95% CI 26 35) in grous randomised to ramiril, 24% (2 28) in those randomised to telmisartan, and 21% (17 25) in those randomised to combination theray (table 3). The risk of develoing new microalbuminuria,, or both, during the trial was not different in grous on telmisartan or ramiril (949, 11 1% 55 Vol 372 August 16, 28

5 vs 118, 11 7%; HR 94, , = 119) but was lower with combination theray than with ramiril (888, 1 4% vs 118, 11 7%; HR 88, 81 96, = 3). Of those with microalbuminuria at baseline, rogression to occurred in 166 (2 12%) on ramiril, 138 (1 77%) on telmisartan, and 125 (1 61%) on combination theray (HR 83, CI , = 114 telmisartan vs ramiril; HR 76, CI 6 96, = 19 combination vs ramiril). Discussion The rimary renal outcome of dialysis, doubling of serum creatinine, and death was similar for telmisartan and ramiril, but was significantly more frequent with the combination of both drugs than with ramiril alone. The comosite of dialysis and doubling of serum creatinine was also more frequent with combination theray. Surrogate renal endoints unexectedly showed contrasting effects, with an adverse effect of combination theray on tyical renal outcomes and on decline in egfr, but showed beneficial effects on roteinuria. Even considering acute and chronic renal failure searately, there was no evidence for a renal benefit with combination theray even though roteinuria was reduced. We rosectively stated that the renal outcomes were the most imortant secondary endoints and a secific statistical analysis lan was develoed in advance. Although ONTARGET was not secifically owered to detect differences of major renal outcomes, nearly 35 rimary renal events were recorded. The comosite of dialysis, doubling of serum creatinine, and death is the commonly used rimary outcome in renal studies. In the ONTARGET study, death was the most common comonent of the comosite rimary outcome. However, there were also several more secific renal outcomes, with 162 cases of dialysis and 461 articiants with a doubling of baseline serum creatinine. Some of the largest revious randomised renal trials, 9,1 with 1513 and 1715 articiants, reorted 696 and 644 rimary comosite renal outcomes of dialysis, doubling of serum creatinine, and death, 347 and 287 cases of dialysis, and 36 and 377 cases of doubling of serum creatinine, resectively. Another study assessed only changes in GFR between an ACE inhibitor and telmisartan and noted no differences between study medications. 11 The resent study therefore suggests that telmisartan has major effects on kidney function that are not materially different from ramiril. Combination theray was not as effective as ramiril, with a greater decline in renal function and more cases of dialysis, doubling of serum creatinine, or both. The loss of renal function of about 1 ml/min er year with combination theray was still lower than in many renal trials of eole at higher renal risk, which tyically reort a loss of 4 1 ml/min er year. 9,1,12 The rate of loss in these studies 9,1,12 varied substantially, from little in individuals A Primary comosite Diabetes No diabetes Overt diabetic nehroathy No overt diabetic nehroathy No diabetes, no hyertension Diabetes or hyertension Microalbuminuria or No microalbuminuria or egfr<6 m 2 /min/1 73 m 2 egfr 6 m 2 /min/1 73 m 2 History of hyertension No history of hyertension B Primary comosite Diabetes No diabetes Overt diabetic nehroathy No overt diabetic nehroathy No diabetes, no hyertension Diabetes or hyertension Microalbuminuria or No microalbuminuria or egfr<6 m 2 /min/1 73 m 2 egfr 6 m 2 /min/1 73 m 2 History of hyertension No history of hyertension Number of atients Number of atients Percentage incidence of rimary utcome in ramiril grou Percentage incidence of Primary outcome in ramiril grou Favours ramiril and telmisartan Favours ramiril with no roteinuria to faster rates in individuals with nehrotic range roteinuria. An initial, short-term decrease in GFR after starting treatment with ACE inhibitors or ARB is exected as a consequence of lowering intraglomerular ressure. 13 This decrease in GFR might even redict a long-term benefit on GFR. 13 Indeed, combination theray in ONTARGET yielded a rather stee initial decrease in GFR; however, Relative risk in telmisartan grou Favours telmisartan Favours ramiril Relative risk in ramril and telmisartan grou for interaction for interaction Figure 4: Relative risk for rimary renal outcome in subgrous (A) Comarison of ramiril and telmisartan. (B) Comarison of ramiril and telmisartan lus ramiril. The interaction term tests whether there is an interaction of the subgrou with the two treatment modalities. Vol 372 August 16,

6 Ramiril gmean gmean Ramiril+telmisartan gmean vs ramiril UACR, Baseline 81 ( 78 84) 83 ( 8 86) 81 ( 78 84) year ratio to baseline 1 17 ( ) 1 8 ( ) 1 5 ( ) 13 < 1 Final ratio to baseline 1 32 ( ) 1 25 ( ) 1 22 ( ) LO ratio to baseline 1 31 ( ) 1 24 ( ) 1 21 ( ) ramiril vs ramiril UACR=urine albumin to creatinine ratio (mg/mmol); Final=study end. gmean=geometric mean. LO=last observation value; for atients with at least one follow-u value, changes from baseline to last observation were comared between grous. All UACR values were log-transformed before analyses; gmean-values are back-transformed. Differences were calculated using an ANOVA model adjusted for baseline values. Number of articiants with measurements=21 76 at baseline, at 2 years, at study end. Table 3: Changes in log urine albumin to creatinine ratio during long-term follow-u GFR did not, as exected, lateau on combination theray but continued to decline faster than with ramiril. A substantial roortion of dialysis events in ONTARGET were acute dialysis. In revious trials that assessed renal outcomes in atients with re-existing renal disease, dialysis events were resumably chronic dialysis, but most studies did not searately reort the effect on acute or chronic dialysis. 14 The rate of chronic dialysis was not significantly different between the three grous in our trial and there was no evidence for articiants on combination theray to have less end-stage renal disease. Doubling of serum creatinine in some renal outcome trials 14 was confirmed by a second measurement of serum creatinine. There was no such confirmatory measurement in ONTARGET, although we believe this would not bias the comarison between the grous. Inhibition of the renin-angiotensin system by ACE inhibitors or ARB reserves renal function better than other antihyertensive agents, secifically in eole with roteinuria above 1 g er day. The greater the roteinuria at baseline (or during follow-u), the bigger the effect of ACE inhibitors comared with other antihyertensive drugs in reducing end stage renal disease. 12 In eole with low-grade roteinuria or unknown urinary rotein excretion, inhibitors of the renin system might be of little benefit in reducing end-stage renal disease. 14,15 In eole with diabetes and microalbuminuria, inhibitors of the renin system revent rogression to ; 2,3 whether this beneficial effect on urinary albumin excretion leads to less end-stage renal disease is unknown. Most articiants in ONTARGET did not have microalbuminuria or. The observation that combination theray was associated with more renal outcomes and a faster decrease in GFR than on ramiril alone is of concern. There was also a benefit of combination theray on indices of roteinuria comared with ramiril alone. These contrasting effects on GFR and on roteinuria were unexected, and more study is needed. 3 Whether modifying low levels of roteinuria attenuates rogression of renal disease has not been investigated in aroriately owered trials. This question is imortant, since many atients with rogressive renal insufficiency have only low-grade roteinuria. The data for loss of egfr with combination theray are therefore clinically relevant. If combination theray had no benefits in the ONTARGET oulation in general, how were results in those at substantial renal risk? Recent meta-analyses have shown that combination theray of ACE inhibitors with ARB reduces blood ressure and roteinuria more than either tye of drug alone. 2,16 There is also a randomised double-blind trial that reorted imroved renal outcomes, such as dialysis, with combination theray as comared with an ACE inhibitor or an angiotensin recetor blocker, 17 although serious inconsistencies in this trial have recently been emhasised. 18 We had exected that combination theray would result in a substantial renal benefit in subgrous of articiants of ONTARGET who had at entry into the study (more than 9). The oint estimate for the rimary outcome in those with overt diabetic nehroathy (more than 7) was in favour of combination theray, a relative risk reduction by 8% with a CI including a risk reduction of 24%, but it was not significant. The interaction term of diabetic nehroathy with effect of combination theray on the rimary outcome was also not significant. Similarly, no benefit of combination theray on the rimary renal outcome was seen in articiants with either microalbuminuria or macro albuminuria, whereas those without microalbuminuria or were at increased risk comared with ramiril theray. Therefore, there is no evidence to suort the use of combination theray in any subgrou of atients included in ONTARGET beyond lowering of urinary albumin excretion. Our data do not necessarily invalidate the hyothesis that roteinuria damages the kidney ie, that lowering roteinuria is beneficial in reventing end-stage renal disease 1,3 or counter the ossibility of a otential benefit of combination theray on renal function in those with. However, our data suggest that roteinuria by itself cannot be taken as a definitive marker of imroved renal function, and that the benefits of any treatment, including combination blockade of the reninangiotensin system on major renal outcomes, remain to be shown. Also, interventions that reduce rotein uria, such as non-steroidal anti-inflammatory drugs, might not effectively slow a decline in renal function Vol 372 August 16, 28

7 Why clinical renal adverse events were seen more frequently with combination theray than with ramiril is unclear. At randomisation, about two-thirds of articiants reorted a history of arterial hyertension, and mean systolic blood ressure was about 142 mm Hg at randomisation and decreased to the mid-13 mm Hg range on randomised treatments during the trial. Average mean systolic blood ressures were by only 2 3 mm Hg lower on combination theray than on ramiril alone. 6 Extended hyotensive eisodes could have contributed to the adverse renal events because articiants of the ONTARGET study were recruited from a oulation with resumably a reduced autoregulatory caacity by the kidneys. Also, eole with unknown renal artery stenosis might be more likely to show a greater decrease in GFR with combination theray than with ACE inhibitors, although there are few reliable clinical data that address this asect. Moreover, visualising the renal vasculature in most atients might not be ractical before starting treatment. The resent analysis has some limitations. First, only a subgrou had overt diabetic nehroathy at baseline, the hallmark of risk for a substantial continuous decline in GFR. The occurrence of dialysis and of doubling of serum creatinine was low. The effect estimates for the renal events had, therefore, wide CIs and were generally of borderline significance. Nevertheless, the number of rimary renal outcomes is larger than numbers reorted in several revious major renal trials. 9,1,12 Serum creatinine was measured locally, and the local methods were not calibrated to a standard. 8 However, the lack of calibration would be a random error and not systematically bias the results between the randomised grous. The data refer to a oulation that received standard cardiovascular theray in a high roortion. 6,7 The effects of combination theray of an ACE inhibitor with an ARB might be different in a less well-treated oulation. Also ossible is that combination theray revents rogressive renal failure and dialysis better than ACE inhibitors or ARB alone in eole with secific roteinuric renal diseases. Contributors All authors were members of the trials steering committee and contributed to the discussions and interretation of the data, and to the writing of the reort. The analysis was lanned by JM, RS, HS, KT, and SY. Data were analysed by LD, JP, HS, and JM. Urine was analysed by MM and XW. All authors had full access to data. No medical writer or other eole were involved in the design, analysis, or writing of this manuscrit. A full list of all investigators has been ublished elsewhere. 6 Conflict of interest statement JM, RS, KT, SY, AB, MK, KM, and LP received consulting and lecture fees and research grants from Boehringer Ingelheim and from other comanies manufacturing angiotensin recetor blockers. HS being an emloyee of Boehringer Ingelheim. References 1 Hilgers KF, Mann JF. ACE inhibitors versus AT(1) recetor antagonists in atients with chronic renal disease. J Am Soc Nehrol 22; 13: Kunz R, Friedrich C, Wolbers M, Mann JF. Meta-analysis: effect of monotheray and combination theray with inhibitors of the renin angiotensin system on roteinuria in renal disease. Ann Intern Med 28; 148: Remuzzi G, Benigni A, Remuzzi A. Mechanisms of rogression and regression of renal lesions of chronic nehroathies and diabetes. J Clin Invest 26; 116: Gerstein HC, Mann JFE, Qilong Y, et al. Albuminuria and cardiovascular events, death and heart failure in diabetic and non-diabetic individuals. J Am Med Ass 21; 286: Phillis CO, Kashani A, Ko DK, Francis G, Krumholz HM. Adverse effects of combination angiotensin II recetor blockers lus angiotensin-converting enzyme inhibitors for left ventricular dysfunction: a quantitative review of data from randomized clinical trials. Arch Intern Med 27; 167: ONTARGET investigators., ramiril, or both in atients at high risk for vascular events. N Engl J Med 28; 358: Teo KK, Yusuf S, Sleight P, et al. Rationale, design, and baseline characteristics of 2 large, simle, randomized trials evaluating telmisartan, ramiril, and their combination in high-risk atients: the Ongoing Alone and in Combination with Ramiril Global Endoint Trial/ Randomized Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease (ONTARGET/TRANSCEND) trials. Am Heart J 24; 148: Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function measured and estimated glomerular filtration rate. N Engl J Med 26; 354: Brenner BM, Cooer ME, de Zeeuw D, et al. Effects of losartan on renal and cardiovascular outcomes in atients with tye 2 diabetes and nehroathy. N Engl J Med 21; 345: Lewis EJ, Hunsicker LG, Clarke WR, et al. Renorotective effect of the angiotensin-recetor antagonist irbesartan in atients with nehroathy due to tye 2 diabetes. N Engl J Med 21; 345: Barnett AH, Bain SC, Bouter P, et al. Angiotensin-recetor blockade versus converting-enzyme inhibition in tye 2 diabetes and nehroathy. N Engl J Med 24; 351: Jafar TH, Stark PC, Schmid CH, et al. Progression of chronic kidney disease: the role of blood ressure control, roteinuria, and angiotensin-converting enzyme inhibition: a atient-level meta-analysis. Ann Intern Med 23; 139: Bakris GL, Weir MR. Angiotensin-converting enzyme inhibitor-associated elevations in serum creatinine: is this a cause for concern? Arch Int Med 2; 16: Casas JP, Chua W, Loukogeorgakis S, et al. Effect of inhibitors of the renin-angiotensin system and other antihyertensive drugs on renal outcomes: systematic review and meta-analysis. Lancet 25; 366: Mann JF, Gerstein HC, Yi QL, et al. Progression of renal insufficiency in tye 2 diabetes with and without microalbuminuria: results of the Heart Outcomes and Prevention Evaluation (HOPE) randomized study. Am J Kidney Dis 23; 42: Matchar DB, McCrory DC, Orlando DC, et al. Systematic review: comarative effectiveness of angiotensin-converting enzyme inhibitors and angiotensin II recetor blockers for treating essential hyertension. Ann Intern Med 28; 148: Nakao N, Yoshimura A, Morita H, Takada M, Kayano T, Ideura T. Combination treatment of angiotensin-ii recetor blocker and angiotensin-converting-enzyme inhibitor in non-diabetic renal disease (COOPERATE): a randomised controlled trial. Lancet 23; 361: Kunz R, Wolbers M, Glass T, Mann JF. The COOPERATE trial: a letter of concern. Lancet 28; 371: Vol 372 August 16,