| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
University Department of Medicine and Therapeutics (J.R.P., M.S., J.M.C.C., H.L.E.), Western Infirmary, West Glasgow Hospitals University NHS Trust, Glasgow G11 6NT, United Kingdom; Department of Medicine (A.D.M.), Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 954, United Kingdom; Second Department of Medicine (S.U.), Yokohama City University School of Medicine, 3-9, Fuku-ura, Kanazawa-ku, Yokohama 236, Japan; and School of Medical and Surgical Sciences (R.D.), University of Nottingham, Derbyshire Royal Infirmary, Derby DE1 2Q4 United Kingdom
Address correspondence and requests for reprints to: Dr. John R. Petrie, University Department of Medicine and Therapeutics, Western Infirmary, West Glasgow Hospitals University NHS Trust, Glasgow G11 6NT, United Kingdom. E-mail: jrp1s{at}clinmed.gla.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Influential reports of improved insulin sensitivity (M) during ACE inhibition (17) have led to discussion of a variety of mechanisms for this putative effect (18). However, many of these reports are based on uncontrolled and/or flawed study designs, indirect measures of M, or studies conducted in subjects receiving potentially confounding medication (19).
The landmark UK–Prospective Diabetes Study has recently emphasized the benefits of antihypertensive treatment in type 2 diabetes, reporting similar efficacy of captopril and atenolol on diabetic complications in these patients (20). Furthermore, a recent study using the euglycemic clamp technique has demonstrated that captopril does not affect M in nondiabetic patients with essential hypertension (21). We now report the results of a double-blind, placebo-controlled crossover trial examining the effect of the long-acting ACE inhibitor trandolapril on M in patients with hypertension and type 2 diabetes.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
). A 2-week washout period was
intercalated between the crossover phases. It was a condition for
continuation in the trial that BP must continue to meet entry criteria
at the end of the placebo run-in.
|
At screening, BP and heart rate were measured after 10 min of supine rest by an oscillometric technique using a Dinamap Critikon semiautomatic sphygmomanometer (Johnson & Johnson Professional Products Ltd., Ascot, Berkshire, UK) maintained and calibrated at regular intervals by the Department of Clinical Physics, Western Infirmary. Appropriately sized BP cuffs were available at all assessments to comply with the recommendations of the British Hypertension Society (22).
Each patient attended three 5-h study mornings (after the placebo run-in phase and at the end of each crossover period) to evaluate the metabolic effects of trandolapril vs. placebo. On each occasion, following an overnight fast, patients traveled to the Clinical Investigation and Research Unit by taxi at approximately 0745 h. After 20 min of supine rest, baseline BP and heart rate were remeasured as above (i.e. 24 h after the last dose of trandolapril or placebo). Baseline blood samples were withdrawn, and trial medication was administered orally with 100 mL water. Sixty minutes later, assessment of whole-body M was commenced using a modification (23) of the hyperinsulinemic euglycemic glucose clamp method of DeFronzo et al. (24). In brief, a primed constant-rate infusion of soluble insulin (1.5 mU/kg·min, Actrapid; Novo Nordisk, Bagsvaerd, Denmark) in a 10% (vol/vol) solution of the patient’s own blood in saline (0.9% sodium chloride; Baxter Healthcare, Norfolk, UK) was administered from 0–180 min. Arterialized serum glucose concentration was measured at the bedside every 5 min. A variable rate infusion of exogenous 20% dextrose (Baxter Healthcare) was given from 2–180 min using an IMED iv infusion system (IMED, Abingdon, UK) to maintain serum glucose at 5.2 mmol/L. Systolic and diastolic BP and heart rate were recorded at 15-min intervals during the clamp. At baseline, and at 60-min intervals during the 3-h procedure, additional blood samples were collected for measurements of serum insulin, C-peptide, serum potassium, triglyceride, plasma noradrenaline, PRA, plasma aldosterone concentrations, and plasma ACE activity. A light meal was provided at the end of each assessment before transport home.
Glucose concentrations were measured at the bedside using a Beckman Coulter, Inc. II Glucose Analyzer (Beckman Coulter, Inc., Fullerton, CA). Triglycerides were measured using an enzymatic method (Merck Diagnostics, Darmstadt, Germany). All blood samples for hormone concentrations were collected in chilled tubes and separated for storage at -20 C or -70 C, as appropriate, until assay. Serum insulin and C-peptide (INCSTAR Corp., Stillwater, Minnesota), plasma aldosterone (Biodata, Milan, Italy), and PRA (Biodata) were measured in batches by direct RIA (intra-assay coefficients of variation, 7%, 7%, 5%, and 5%, respectively). Plasma noradrenaline was assayed by high-performance liquid chromotography with electrochemical detection (intra-assay coefficient of variation, 15%). Plasma ACE activity (EU/L) was determined from the rate of hippuric acid release from an artificial substrate of angiotensin I (intra-assay coefficients of variation, 2%).
The distribution of all data was checked using the Shapiro-Wilks test, and log transformation was performed where appropriate (Minitab statistical package; Minitab Inc., State College, PA). Results are accordingly expressed either as arithmetic means (±SD) or geometric means. Unless otherwise stated, "placebo" refers to the crossover rather than the run-in placebo period. M in mg(glucose)/kg·min was calculated by applying the space correction of DeFronzo et al. (24) to the glucose infusion rate under steady-state conditions during the final 40 min of each clamp. The method of Hills and Armitage (25) was used to examine data for a period effect or a treatment-period interaction (carryover effect). Thereafter, the treatment effect was examined by comparing M at the end of each crossover phase using a paired t test; 95% confidence intervals (CIs) are quoted. Mean systolic BP, diastolic BP, and heart rate during each clamp study for each individual patient was used as a summary measure (26) for comparison between study days (paired t test). Insulin, C-peptide, potassium, triglycerides, ACE activity, renin, aldosterone, and noradrenaline profiles were compared between study days by three-way ANOVA; Bonferroni-corrected post-hoc t tests were performed at individual time points.
The table summarizing previous randomized placebo-controlled trials was compiled using searches of MEDLINE and EMBASE from 1980 onward using the keywords "angiotensin-converting enzyme inhibitor," "diabetes," "dipeptidyl carboxypeptidase inhibitor," "hypertension," "insulin," and "insulin sensitivity." Further references were obtained by manually searching for references cited in the papers retrieved. Two crossover studies that did not incorporate a placebo comparison but were otherwise rigorously designed (27, 28) were not excluded because they were judged to be at least as well-designed as the reference parallel group study (17).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The euglycemic clamp procedure was well tolerated by all patients, except for one who had a hypotensive episode during his second study day, and another who had hypoglycemic symptoms at a blood glucose concentration of 5.4 mmol/L. Both patients were withdrawn from the trial, and their data were excluded. Tablet counts suggested that over 95% of trial medication was ingested as prescribed.
Plasma glucose and whole body M
There were no significant differences in plasma glucose
concentrations during treatment with trandolapril when compared with
placebo; mean ± SD fasting plasma glucose
concentrations were 6.3 ± 1.31, 6.0 ± 0.91, and 6.1 ±
1.13 mmol/L for placebo run-in, placebo, and trandolapril,
respectively. The corresponding plasma glucose concentrations for the
last 40 min of each clamp were 5.1 ± 0.29, 5.1 ± 0.30, and
5.3 ± 0.32 mmol/L. There was no evidence of a period effect
(P > 0.1) or a treatment-period interaction
(P > 0.1). There was no significant difference in
whole body insulin-stimulated glucose uptake (M) between the three
study days (Fig. 2): placebo (run-in),
5.2 ± 1.98 mg/kg·min; placebo, 5.3 ± 1.70 mg/kg·min;
trandolapril, 5.1 ± 1.65 mg/kg·min; P = 0.58
(trandolapril vs. placebo); 95% CIs, -0.74 and 0.43
(i.e. the trial had 95% power to exclude a 14% decrease
and an 8% increase in M).
|
Trandolapril caused a significant decrease in trough plasma ACE
activity (i.e. 24 h after the last dose) (Fig. 3). Following drug administration,
trandolapril produced a further prompt decrease in plasma ACE activity.
At the same time, a trend toward a rise in PRA was observed (not
statistically significant). No differences were observed between
trandolapril and placebo with respect to plasma aldosterone and
noradrenaline concentrations either at baseline or during
hyperinsulinemia.
|
Mean BP during the clamp procedure was 14/6 mm Hg lower during
treatment with trandolapril when compared with placebo [systolic,
135 ± 18.1 vs. 149 ± 19.8, P <
0.001 (95% CIs, -16.5 and -12.3); diastolic, 73 ± 13.1
vs. 79 ± 12.5, P < 0.001 (95% CIs,
-8.0 and -5.1])]. Trough BP (i.e. 24 h after the
last dose) during chronic treatment with trandolapril tended to be
lower by 6/4 mm Hg when compared with placebo [systolic, 157 ±
20.4 vs. 163 ± 21.8, P = 0.11 (95%
CIs, -12.5 and 1.4); diastolic, 86 ± 8.7 vs. 90
± 12.7, P = 0.08 (95% CIs, -7.04 and 0.4)]. There
were no statistically significant differences in heart rate among the
three study days (Fig. 4).
|
Serum insulin concentrations increased to a plateau within the
first 60 min of the infusion with no significant differences between
the three study days; C-peptide concentrations tended to suppress to
60% of baseline (Fig. 5). No
differences were observed between trandolapril and placebo with respect
to serum potassium and triglyceride concentrations either at baseline
or during hyperinsulinemia.
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
) (17, 30, 31, 32, 33, 34, 35, 36), only one other trial
published very recently (21) has incorporated all of the following
features: 1) a double-blind, placebo-controlled, crossover design; 2)
assessment of M using a highly reproducible technique; 3) adequate
power to avoid clinically important type 2 error; and 4) reliable
exclusion of potentially confounding carryover effects. Only the
present trial was conducted in patients with both hypertension and
diabetes.
|
) (27, 28, 42, 43, 44, 45, 46, 47), but, like the "positive" trials, few of
these "negative" studies have been entirely satisfactory in terms
of methodology. In trying to make sense of the literature, we should first discount early studies that relied on indirect measurements of M (48). Second, it is self-evident that uncontrolled studies (41) and assessments of M "before and after" therapy are not definitive (35, 41, 47, 49). Third, open and single-blind clamp studies (27, 47) are open to investigator bias (50). Finally, it is important to consider study design (parallel group vs. crossover) because a higher proportion of the former type of studies have reported an insulin-sensitizing effect of ACE inhibition.
Use of a parallel group design (17, 33, 34, 45, 46) might be thought a priori to be sufficient in this setting: however, careful review of previous literature reveals shortcomings in trials of this nature. Measurement of M is relatively invasive and time-consuming; thus, studies are usually conducted on small numbers of subjects. As M varies 3-fold even in healthy individuals (51), parallel treatment groups may be poorly matched at baseline, leading to type 1 error. Indeed, presumably for this reason, all of the "parallel group" trials report statistical comparisons with baseline (run-in) measurements rather than direct comparisons with data from the control group (17, 33, 34, 45, 46). Failure to make full use of placebo control data leaves the analysis vulnerable to effects resulting from unmeasured nonexperimental factors, particularly withdrawal of previous therapy with thiazide diuretics (52). For these reasons, it is clear that in relatively small numbers of patients a placebo-controlled crossover design, in which each patient acts as his or her own control, avoids many of these pitfalls, particularly if absence of period and carryover effects can be demonstrated.
One of the most widely cited and well-publicized studies in this area reported an improvement in M in nondiabetic patients with hypertension randomized to captopril treatment vs. hydrochlorothiazide (17). The original crossover aimed to compare captopril with hydrochlorthiazide and did not include a direct comparison with placebo. However, data in the second treatment period were unsuitable for analysis owing to a carryover effect, and results were presented separately for the parallel groups as comparisons with the baseline placebo period. The captopril group (n = 23) at the start of treatment had similar measured M to that of the diuretic-treated group (n = 27) at the end of treatment: this suggests that the reported treatment effect may simply reflect regression toward the mean. Despite these shortcomings, this trial has been extremely influential in support of the perception that ACE inhibitors improve M.
More than half of the patients in the present crossover trial had not previously taken antihypertensive therapy; of those previously treated, none had been on thiazide diuretics and the remainder had been on monotherapy, which had been withdrawn for at least 4 weeks before the first clamp study. Lack of a period effect or a carryover effect was confirmed by the Hills-Armitage method, and the observed similarity between placebo run-in data and placebo crossover data confirmed adequacy of the washout period. The three previous crossover trials that have reported an insulin-sensitizing effect (30, 31, 32) are characterized by the absence of an explicit analysis to exclude period and carryover effects. In addition, two had washout periods of 1 week or less (30, 32) and the other was not double-blind (31).
For these reasons, we believe that the results of the present trial in patients with diabetes, together with those of the recently published study by Wiggam et al. (21) in nondiabetic patients, are more robust than data from previous trials that have examined the potential insulin-sensitizing effect of ACE inhibition. In addition, we suggest that previous positive reports are attributable to deficiencies in study design. However, two explanations other than study design could be advanced. First, glucose metabolism may respond differently to ACE inhibition in different subgroups of patients. This does not seem to be the most likely explanation for the present discrepancy because improvements in M have previously been reported in both hypertensive patients with normal glucose tolerance (17) and in diabetic hypertensive patients (30, 53). Second, it is possible that the putative insulin-sensitizing effect is drug-specific rather than class- specific: some investigators have invoked a specific effect of the sulphydryl group of captopril (36) (although this seems unlikely in the context of the study by Wiggam et al. (21), which was able to exclude a 10% increase in M in patients treated with captopril).
Other lines of evidence have reinforced the widespread belief that ACE inhibitors improve M. In a retrospective cohort study, Herings et al. (54) reported an excess of hypoglycemic episodes among diabetic patients treated with ACE inhibitors, a finding that has been independently replicated in a different population (55) but may be attributable to unmeasured confounding variables (56). In addition, patients treated with captopril in the Captopril Prevention Project had a lower risk of developing type 2 diabetes mellitus than those on conventional therapy (57). However, the randomization process was flawed in this trial, and it is noteworthy that no differences in HbA1c concentrations were observed in patients with type 2 diabetes randomized to atenolol or captopril in the UK–Prospective Diabetes Study (20) nor between patients with type 1 diabetes treated with lisinopril or placebo in the EUCLID study (4).
In summary, we report the first double-blind, placebo-controlled crossover trial adequately to evaluate the effects of ACE inhibitor therapy on M in patients with hypertension and type 2 diabetes. Our findings are consistent with recent evidence in nondiabetic patients (21). Despite conflicting evidence from previous poorly controlled but well-publicized studies (17), effective antihypertensive treatment with a long-acting ACE inhibitor did not improve whole body M. Although ACE inhibitors are known to be highly effective in a number of therapeutic contexts, putative beneficial effects on glucose metabolism should not be invoked as a justification for their use as first-line antihypertensive therapy in patients with type 2 diabetes in the absence of other specific indications.
|
Footnotes |
|---|
Received August 4, 1999.
Revised November 2, 1999.
Accepted February 8, 2000.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. K McGuire, J. R Winterfield, J. A Rytlewski, and E. Ferrannini Blocking the renin-angiotensin-aldosterone system to prevent diabetes mellitus Diabetes and Vascular Disease Research, March 1, 2008; 5(1): 59 - 66. [Abstract] [PDF]
|
|
|
|
 |
|
 G. Bakris, M. Molitch, A. Hewkin, M. Kipnes, P. Sarafidis, K. Fakouhi, P. Bacher, J. Sowers, and on behalf of the STAR Investigators Differences in Glucose Tolerance Between Fixed-Dose Antihypertensive Drug Combinations in People With Metabolic Syndrome Diabetes Care, December 1, 2006; 29(12): 2592 - 2597. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. S. Hermann, W. Li, H. Dominguez, N. Ihlemann, C. Rask-Madsen, A. Major-Pedersen, D. B. Nielsen, K. W. Hansen, M. Hawkins, L. Kober, et al. Quinapril Treatment Increases Insulin-Stimulated Endothelial Function and Adiponectin Gene Expression in Patients with Type 2 Diabetes J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 1001 - 1008. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. E. Kershaw and J. S. Flier Adipose Tissue as an Endocrine Organ J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2548 - 2556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Davies and A. Struthers Review: The potential benefits of aldosterone antagonism in Type 2 diabetes mellitus Journal of Renin-Angiotensin-Aldosterone System, September 1, 2002; 3(3): 150 - 155. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Euglycemic
clamp study.
, Placebo run-in; 
, placebo; 
,
trandolapril; **, P < 0.01 for trandolapril
vs. placebo.
