NDT Advance Access originally published online on June 17, 2006
Nephrology Dialysis Transplantation 2006 21(9):2352-2353; doi:10.1093/ndt/gfl316
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Progression of renal disease—can we forget about inhibition of the renin–angiotensin system? Authors’ reply
1 Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine and 2 Centre for Clinical Pharmacology, Department of Medicine, BHF Laboratories, University College London, UK
Correspondence and offprint requests to: Dr Juan P. Casas, MD, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, UK. Email: juan.pablo-casas{at}lshtm.ac.uk
Keywords: ACE inhibitors; angiotensin receptor antagonists; antihypertensive drugs; clinical trials; end-stage renal disease
In their editorial in the present issue, we note that Mann et al. have managed, for the second time, to overcome their stated reluctance to criticize others. In the current piece, they raise issues that they aired previously in a letter to the Lancet [1]. We will respond again to the arguments they raise, and would also like to make readers aware of some relevant points, so that they can put the methodological criticisms of our paper into context. Our work was submitted to the Lancet, where it underwent internal review, then external peer review by three experts, followed by a statistical review. No significant concerns were raised. Like all articles, ours may have flaws but we doubt that these are such as to invalidate the work, as suggested by Mann and colleagues. While we respect their right to hold a contrary ‘expert’ view to the findings of our research, we are concerned that they purport to present their view as evidence-based. We deal point-by-point with their comments and hope that future publications in this area will be based on original research or rigorous systematic reviews of such research, rather than opinion.
There are a number of inaccuracies in the re-analysis of our work by Mann et al. Firstly, doubling of serum creatinine and incidence of end-stage renal disease (ESRD) is not the most frequently used primary renal outcome. As is clear from our analysis [2], albuminuria is the most commonly measured outcome, at best a surrogate for more clinically important endpoints. We are not aware of any publication that reports on the conditional endpoint of ‘doubling of serum creatinine and ESRD’ (which would be uncommon), whereas we have documented in detail the number of clinical trials that reported on doubling of serum creatinine or ESRD. If Mann et al. mean that we did not present data on the composite endpoint of doubling of serum creatinine or ESRD, we should point out that such composite endpoints are constructed to enhance the power of a study recording rare endpoints; our meta-analysis of all available data enabled us to overcome this limitation. Secondly, the treatment effect summary from a meta-analysis does not simply involve adding up all the events and individuals and treating them as a single large study, without considering that they arise from different trials with differences in sample size. The ‘naïve calculation’ (to use their own words) presented by Mann et al. showing a relative risk for doubling of serum creatinine of 0.7 is incorrect. Lastly, we were cautious in our interpretation of the positive findings of our analysis (such as the reduction in ESRD in non-diabetic renal disease) because there was evidence of publication bias.
We evaluated all available data using a systematic review. The previous participant-level meta-analysis to which Mann et al. refer (with 176 ESRD events) provides information from a selected subset of all available studies: those who agreed to provide individual participant data [3]. These form about one-sixth of all available data on ESRD in our analysis (1536 events). Mann et al. suggest the need to ‘adjust for’ potential confounding factors such as smoking or age as can be done with access to patient-level data. The whole point of randomization is to ensure that known and unknown confounding factors are equally balanced across the comparison groups. While such adjustment might be appropriate in observational studies, we only included randomized trials in our analyses. The second meta-analysis referred to by Mann et al. includes data on albuminuria from 330 patients that were not included in our analysis, because these studies did not meet our entry criteria (follow-up too short, randomization to angiotensin-converting enzyme (ACE) inhibitors or angiotensin-II receptor blocker's (ARB's) alone or no relevant renal outcomes evaluated).
Are Mann et al. stating that they do not believe that blood pressure (BP) reduction per se is renoprotective? This would seem to be an unusual position to adopt, and one that does not seem to be shared by triallists and meta-analysts, who have gone to great lengths to control for the effects of BP differences on renal outcomes in ACE-I/ARB trials. It is difficult to document BP difference accurately and repeatedly enough to eliminate residual confounding. Attempts to take account of BP in this way, even using data from the substantially larger and better powered trials of BP-lowering agents in prevention of cardiovascular outcomes, has been met with scepticism. For this reason, we asked to what extent BP differences in ACE-I/ARB trials explained the observed renoprotection, hypothesizing that the treatment effects of ACE-I/ARB would be greater in placebo-controlled trials (which was the case) and that meta-regression would identify an association between BP differences and outcomes (which was also the case). These observations are concordant with epidemiological data on the effects of BP on renal disease progression, and are consistent with the much greater body of evidence provided by trials measuring cardiovascular endpoints. Indeed, in Jafar et al.'s [5] later publication, BP was an extremely important determinant of renal disease progression [5].
Mann et al. question the inclusion in our meta-analysis of data on renal outcomes from Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trail (ALLHAT), arguing that ALLHAT was unable to provide information on renal outcomes. However, they acknowledge that ALLHAT was the largest ever randomized clinical trial comparing different antihypertensives undertaken in the United States. Renal outcomes were specified a priori in the ALLHAT protocol [6]. Thus, ALLHAT was the only large, high-quality randomized trial with pre-specified renal outcomes. Mann et al. also suggest that the ALLHAT population is not representative of patients at risk of diabetic renal disease because of the low incidence of nephropathy in these patients. However, the large sample size in ALLHAT meant that it provided one-third of all end-stage renal disease events. Moreover, in a post hoc analysis by category of glomerular filtration rate at baseline, there was no signal for a benefit in higher risk patients allocated to the ACE-I arm. Some people may not like the findings of ALLHAT, but this does not mean the results can be dismissed as untrue.
Systematic reviews and meta-analysis that summarize the totality of available evidence are likely to provide the best guide to decision-making. An alternative reasoned argument has been made that, when available, we should rely on the findings of a single, large high-quality randomized trial rather than a meta-analysis of all studies [7]. Mann et al. suggest a third novel strategy: to rely neither on a summary of all studies nor on the results of the single largest trial, but instead to base our opinion on a selected subset of smaller studies. We believe such a strategy is flawed. When assessing the effects of inhibitors of the renin–angiotensin system on renal outcomes, whether one chooses to believe the single largest study or a pooled summary of all studies, uncertainty remains about a specific renoprotective effect of these drugs over and above BP lowering.
 |
Acknowledgments |
|---|
A.D.H. acknowledges support of the British Heart Foundation. L.S. acknowledges support of the United Kingdom Medical Research Council.
Conflict of interest statement. A.D.H is a member of the Editorial Board of the Consumers’ Association Drug and Therapeutics Bulletin. The other authors have declared no conflicts of interests.
|
References |
|---|
 Top References |
|---|
- Casas JP, Vallance P, Smeeth L, Hingorani AD, MacAllister RJ. Renoprotective effects of renin-angiotensin-system inhibitors: authors reply. Lancet 2006; 367: 900–902[Web of Science][Medline]
- Casas JP, Chua W, Loukogeorgakis S et al. Effect of inhibitors of the renin–angiotensin system and other antihypertensive drugs on renal outcomes: systematic review and meta-analysis. Lancet 2005; 366: 2026–2033[CrossRef][Web of Science][Medline]
- Jafar TH, Schmid CH, Landa M et al. ACE inhibitors and progression of non-diabetic renal disease: a meta-analysis of patient-level data. Ann Int Med 2001; 135: 78–87
- Strippoli GF, Craig M, Schena FP, Craig JC. Antihypertensive agents for primary prevention of diabetic nephropathy. J Am Soc Nephrol 2005; 16: 3081–3091
[Abstract/Free Full Text] - Jafar TH, Stark PC, Schmid CH et al. Progression of kidney disease: the role of blood pressure control, proteinuria, and ACE inhibition: a patient-level metaanalysis. Ann Int Med 2003; 135: 73–87
- Davis BR, Cutler JA, Gordon DJ et al. Rationale and design for the Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial (ALLHAT). ALLHAT Research Group. Am J Hypertens 1996; 9: 342–360[CrossRef][Web of Science][Medline]
- Bailar JC,3rd. The promise and problems of meta-analysis. N Engl J Med 1997; 337: 559–561
[Free Full Text]
Accepted in revised form: 2. 5.06
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||