Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

ACE-inhibitors for renal disease and diabetes?

Chronic renal disease [1]
Diabetes [3]
Type 1 diabetes [4]

Bandolier has heard it said that most patients with chronic renal disease or diabetes should be taking ACE inhibitors. Evidence seems to be accumulating, in the shape of systematic reviews and large randomised trials. A brief review, then, of some of the evidence on outcomes, seems in order.

Chronic renal disease [1]

This review used a MEDLINE search (to June 1999) for studies of ACE inhibitors in chronic renal disease, supplemented with hand searching of abstracts, examination of reference lists of reviews, and solicitation of pharmaceutical companies. A broad definition of chronic renal failure was used, consisting of albuminuria greater than 30 mg/day (21 µg/minute) and/or elevated serum creatinine (more than 106 µmol/L for women and 124 µmol/L for men).

Trials were eligible if they were randomised, had a parallel comparison of ACE inhibitor with placebo, had a minimum follow-up of one year and studied adults. Information unavailable in the published reports was requested from the original authors.


There were nine studies with 650 patients with chronic renal failure and microalbuminuria initially, and who were treated with ACE inhibitor for one to five years (average three years). ACE inhibitors were enalapril 10 or 20 mg, captopril 100 mg and lisinopril 10 mg. The outcome was development of macroalbuminuria. The risk of developing macroalbuminuria was lower with ACE inhibitor (10%) than with placebo (24%) (Figure 1; Table 1). Treating seven patients with chronic renal failure and microalbuminuria for three years with an ACE inhibitor would prevent one of them developing macroalbuminuria.

Figure 1: Developing macroalbuminuria

Table 1: Chronic renal disease

      Number/total (%)    
Patients at onset Outcome Number of trials ACE inhibitor Placebo Relative risk (95% CI) NNT (95% CI)
Microalbuminuria Developed macroalbuminuria over 3 years 9 30/326 (10) 74/316 (24) 0.4 (0.3 to 0.6) 7 (5 to 12)
Overt proteinuria End-stage renal disease or doubled creatinine over 2 years 7 115/700 (17) 194/689 (28) 0.6 (0.5 to 0.7) 9 (6 to 14)

There were seven studies with 1,400 patients with chronic renal failure and overt proteinuria initially, and with ACE inhibitor for one to three years (average two years). ACE inhibitors were enalapril 5-40 mg, captopril 75 mg and ramipril 2.5-5 mg and benazepril 10 mg. The outcome was development of end stage renal disease or doubling of serum creatinine. The risk of developing this outcome was lower with ACE inhibitor (17%) than with placebo (28%) (Figure 2; Table 1). Treating nine patients with chronic renal failure and overt proteinuria for two years with an ACE inhibitor would prevent one of them developing end stage renal disease or doubling their serum creatinine.

Figure 2: Developing end-stage renal disease

Diabetes [3]

The Heart Outcomes Prevention Study (HOPE [2]) was a large (9,000 patient) study examining whether the addition of an ACE inhibitor to current medical regimen of patients with vascular disease or diabetes can lower the risk of cardiovascular events. It showed that 10 mg daily ramipril in high risk patients over five years reduced the risk of myocardial infarction, stroke, or cardiovascular death. Treating 1000 patients with ramipril for four years prevented about 150 events in about 70 patients. A separate analysis showed the effects in the 40% of patients with diabetes [3].

Patients recruited were 55 years or older, who had a history of cardiovascular disease, diabetes, plus at least one other cardiovascular risk factor (total cholesterol above 5.2 mmol/L, HDL cholesterol ≤0.9 mmol/L, hypertension, microalbuminuria, or current smoking). They were randomised to 10 mg ramipril or matched placebo, and were followed for 4.5 years. The main outcome was myocardial infarction, stroke, or cardiovascular death, a combined outcome of bad things. These were also examined individually, with a number of secondary outcomes.


The main results are shown in Table 2, as percentages of events occurring. Those shown were statistically significant. When major cardiovascular and microvascular events were taken into account, 15 high-risk people with diabetes would have to be treated with ramipril for 4.5 years to prevent one of them having a myocardial infarction, stroke, cardiovascular death, admission to hospital for heart failure, a revascularisation procedure, development of overt nephropathy, laser therapy for retinopathy, or renal dialysis.

Table 2: ACE inhibitors for diabetes

  Percent occurring with
Outcome Ramipril Placebo
Number of patients 1808 1769
Combined primary outcome 15.3 19.8
MI 10.2 12.9
Stroke 4.2 6.1
Cardiovascular death 6.2 9.7
Total mortality 10.8 14
Revascularisation 14 16.4
Overt nephropathy 6.5 8.4
Any heart failure 11 13.3
TIA 4.4 5.9

Type 1 diabetes [4]

Should all patients with type 1 diabetes mellitus and microalbuminuria receive ACE inhibitors? This question was answered by a meta-analysis [4] performed on individual patient data. A MEDLINE search supplemented by reference lists and other meta-analyses identified 12 studies meeting inclusion criteria. These were normotensive diabetics with microalbuminuria (20 to 200 µg/minute) treated with ACE inhibitor or placebo in a randomised trials with at least one year follow up. Information about individual patients was requested from original investigators. The primary outcome was the change in albumin excretion rate.


There were 12 trials with 700 patients. Early analysis showed a time-dependent response, and so the analysis was restricted to a two-year follow up from 10 of the 12 studies (646 patients). There was a marked beneficial effect of ACE inhibitors, with albumin excretion rate 50% lower at two years in treated versus untreated patients. The effect was highly dependent on baseline albumin excretion rate, with a small effect (18%) at the lower boundary of the target range (20-200 µg/min) and a higher (74%) reduction at the higher boundary.

Progression to macroalbuminuria was reduced. The odds ratio was 0.4 (0.3 to 0.6). Regression to normal albumin excretion was increased. The odds ratio was 3.1 (2.2 to 4.4).


These three papers show useful clinical effects of ACE inhibitors on renal and cardiovascular outcomes in chronic renal disease and diabetes. This occurs with fixed dosing, rather than with dosing adjusted to provide the maximum blood pressure drop. Though there was some statistically lower blood pressure, it was usually less than 2 mmHg, beyond the ability of most to measure. We probably have to think beyond blood pressure for these effects. The outcomes are important. Preventing bad things happening to high risk renal or diabetic patients is good, and the ACE inhibitors appear to be effective in that.

Interesting were the different ways in which evidence was obtained. We have the contrast between meta-analysis of studies [1], of individual patients from studies forming a subgroup not defined in the original studies themselves [4], and a large subgroup within a larger overall trial [3].

All the papers are naturally full of detailed statistics. That is as it should be. Two also give results in terms our tired brains can handle. Notable is the choice of presenting information in terms of 'all bad things' [3], so we know that whatever bad thing it is, treating 15 diabetic high-risk patients with ramipril for 4.5 years will prevent one of them [3]. Another [1] gives us raw data with the statistics, so we can look for ourselves and do what we want, which is how Bandolier could calculate an NNT.

But the third [4] gives us only odds ratios. It is terrific that ACE inhibitor treatment of type 1 diabetics with microalbuminuria reduces progression of macroalbuminuria and promotes return of normal albumin excretion. What is missing is any suggestion of the therapeutic effort needed to achieve this. Odds ratios just don't do this. Table 3 calculates the odds ratios for a theoretical reduction in progression to macroalbuminuria. Given 700 patients split equally between treatment and placebo, the same odds ratio (0.4) as found in the meta-analysis could be an NNT of 4 or 23. The proportion of patients benefiting could be 25% or 4%. So a plea to the academics. Good statistical methods are vital. Getting the statistical tick is the priority. But when you've got it, please express the result in ways that ordinary mortals can understand and use for their patients. If you can't do that, what's the point?

Table 3: Changing NNT at fixed odds ratio

Treatment (n=350) Placebo (n=350)    
Number % Number % Odds ratio NNT
10 3 25 7 0.41 23
20 6 50 14 0.39 12
40 11 90 26 0.39 7
60 17 120 34 0.41 6
100 29 180 51 0.39 4
150 43 230 66 0.40 4

Adverse events don't receive the detailed treatment they deserve in any of these papers.


  1. AV Kshirsagar et al. Effect of ACE inhibitors in diabetic and nondiabetic chronic renal disease: a systematic overview of randomized placebo-controlled trials. American Journal of Kidney Diseases 2000 35: 695-707.
  2. HOPE Study Investigators. Effects of an angiotensin-converting enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. New England Journal of Medicine 2000 342: 145-153.
  3. HOPE Study Investigators. Effects of ramipril on cardiovascular and microvascular outcomes in people with diabetes mellitus: result of HOPE study and MICRO-HOPE substudy. Lancet 2000 355: 253-259.
  4. ACE Inhibitors in Diabetic Nephropathy Trialist Group. Should all patients with type 1 diabetes mellitus and microalbuminuria receive angiotensin-converting enzyme inhibitors? Annals of Internal Medicine 2001 134: 370-379.
previous or next story in this issue