The roles of antihypertensive agents and in particular those that inhibit the RAS in the acceleration or deceleration of diabetes manifestation have been discussed controversial and study results on this question are not consistent.

The RAS itself plays a pivotal role in the development of diabetes. Over-activity appears to be linked to reduced insulin and glucose delivery to the peripheral skeletal muscle and impaired glucose transport and response to insulin signalling pathways, thus increasing insulin resistance 10 . Activation of a local pancreatic RAS, in particular within the islets, may represent an independent mechanism for the progression of islet cell damage in diabetes. In fact, impaired pancreatic islet function may predominate quantitatively over peripheral insulin resistance in impaired glucose tolerance 11 .

Drugs that interrupt the RAS like angiotensin converting enzyme inhibitors (ACEi) and angiotensin receptors blockers (ARBs) are likely to be beneficial in the prevention of diabetes 10 12 . A series of recent large-scale prospective randomised studies of 3–6 year duration such as CAPP, INSIGHT, LIFE or ALLHAT, reported a remarkably consistent reduction in the incidence of type 2 diabetes in hypertensive patients reported with either ACEi-based or ARB-based therapy (reviewed by Jandeleit-Dahm in 10 ). The comparator groups were based on thiazide-diuretics, β-blockers, the calcium channel blocker amlodipine or placebo, respectively.

In a large meta-analyis, Abuissa et al. calculated the average risk reduction in 6 of these trials using ACEi and 7 trials using ARBs. The reduction of new onset-diabetes was 24% for ACEi, 23% for ARBs and 23% for the combined data-set 13 . Furthermore a recent network meta-analysis of randomized controlled trials showed that while patients taking betablockers and diuretics show an increased incidence of diabetes, it is reduced in patients using ACEi or ARBs (Figure 1) 14 15 .

Endpoint studies to elucidate the role of antihypertensive agents on new onset diabetes related morbidity and mortality are however scarce. The VALUE trial with valsartan was the only trial to include new-onset diabetes as a pre-specified endpoint 16 . Patients were normoglycemic, those with abnormal glucose values were excluded. While 16.4% of patients in the amlodipine arm (up to 10 mg) developed diabetes over a mean follow-up of 4.2 years, 13.1% developed such in the valsartan arm (up to 160 mg); p < 0.0001. The ASCOT-BPLA study, which was a randomised controlled trial of the prevention of CHD and other vascular events by BP and cholesterol lowering in a factorial study design, was prematurely stopped in December 2004 17 18 . The study was designed to resolve whether newer antihypertensive strategies that use calcium channel blockers (CCBs) and ACEi are superior to older treatments using betablockers and diuretics. As a key finding there was a substantial excess of new diabetes (increase of 30%) in the beta blocker/diuretic arm 17 . DREAM investigated the effect of ramipril (up to 15 mg) compared to placebo 6 . In this randomised controlled trial ramipril significantly increased regression to normoglycemia in patients with impaired glucose tolerance. It did however not influence the risk of a combined endpoint consisting of new-onset diabetes or death over a 3 year observational period. Interpretation of DREAM is limited by a number of details: 1) hypertension was not an inclusion criterion (mean blood pressure at baseline 136/83 mmHg), 2) comparison was made to placebo instead of diuretics or betablockers (which would be reasonable based on the analysis of Elliott 15 ) and 3) betablockers were allowed in both the ramipril and placebo groups.

Thus, despite the strong evidence for a reduction of new-onset diabetes from several RCTs and meta-analyses, there is an ongoing controversy about the clinical significance, the comparability of agents within one drug class, or the generalisability of these findings into clinical practice 19 20 21 . The "ACE inhibitor-based versus diuretic-based antihypertensive primary treatment in patients with prediabetes" (ADaPT) study addresses this issue. On the basis of the existing body of evidence, it appeared likely that patients with impaired fasting glucose (IGF) according to the screening on pre-diabetes (PreDiSc Score) will benefit from tight blood pressure control and further effects from RAS-inhibition by the ACEi ramipril in terms of manifestation of type 2 diabetes. The long-term outcomes of this treatment regimen will be compared to a regimen based on diuretics and/or β-blockers.

The ADaPT investigation is a comparative Post Marketing Surveillance according to §67(6) German Drug Law, performed by the German Hypertension League. It was designed as an open, prospective, non-randomised parallel group observational investigation in 150 general practices (general physicians and internists) throughout Germany.

For definitions of normal glucose tolerance (NGT), impaired fasting glucose (IFG), impaired glucose tolerance (IGT) and overt diabetes (DM) see Table 1. PreDiSc Score: The score indicates the presence of pre-diabetes with a diagnostic sensivity of 78% and a specificity of 37% using the following parameters: blood pressure (BP) ≥ 140/90 mmHg, capillary fasting blood glucose ≥ 100 mg/dl (STIX) and age ≥ 55 years. Sensivity can be increased to 79% and specificity to 74% by additional determination of the HbA1c value (≥ 6%) 22 .

Patients in Group 1 receive Ramipril either as monotherapy (Delix^®, Sanofi-Aventis, Berlin) or in combination with Felodipin (Delmuno^®, Sanofi-Aventis, Berlin), patients in Group 2 received any other diuretic-based or β-blocker-based therapy without using ACEi or ARBs. Ramipril was chosen as the antihypertensive drug in one group, as long-standing experience from several clinical studies including large endpoint studies have accumulated with this agent 23 24 25 . The Heart Outcomes Prevention Evaluation (HOPE) Study showed that ramipril is effective in preventing major cardiovascular events in high-risk patients without hypertension or those whose hypertension was sufficiently controlled with other treatments 26 .

Generally, treatment regimens in this study can be chosen in accordance with the recommendations of the German Hypertension League and the European Society of Hypertension 27 28 For initial treatment monotherapy or a low-dose combination regimen is suggested. If the response is inadequate, possible options include increasing the dose, changing the drug or introduction of further combination drugs 29 . All drugs are administered within the approved labelling.

Advice about the investigation has been obtained by the institutional review board of the Charité Berlin, Germany – Campus Benjamin-Franklin. Written informed consent was obtained from every patient in writing. The planned follow-up period is four years.

The primary evaluation criterion of this observational study is the first manifestation of type 2 diabetes according to the current guidelines of the German Diabetes Society (see table 1). 30 Further evaluation criteria are the deterioration of pre-diabetes indicated by an increase of HbA1c of at least 10% compared to baseline value within 4 years, the initiation of antidiabetic glucose lowering medication, an increase of fasting glucose levels, change of HbA1c compared to baseline, BP reduction, achievement of the target BP <130/80 mmHg after 12 months and at the 4 year follow-up, time needed to reach target BP, major cardiovascular (CV) events (first manifestation of symptomatic coronary heart disease (CHD) and/or peripheral arterial occlusive disease and/or cerebrovascular events), type and frequency of Adverse Events (AE) or Serious Adverse Events (SAE), and total mortality. In the diuretic-based therapy group, time to switch to ACE-based or ARB-based therapy will also be analysed.

Table 2 summarizes the investigational plan. Patients are seen at 7 scheduled visits. Vital signs (BP, heart rate) will be complemented by height and weight measurements (determinations of body mass index) and waist and hip circumference measurement at after 3, 6 and 12 months and thereafter at yearly intervals. Further, lab examinations of glucose, lipids, inflammatory (high sensitive C-reactive protein) and renal parameters (with cystatin C to assess renal function 30 31 ) will be done in the same intervals. Ambulatory BP monitoring is facultative and will be performed in a subset of patients. AEs will be recorded and their severity, course and relation to the medication assessed by the treating physician. Prior as well as concomitant diseases and concomitant medication will also be assessed.

Sample size calculation for the primary endpoint was made under the assumption – based on the results of the ALLHAT study – that during the 4-year observational period 10.0% of patients in the ramipril-based antihypertensive regimen and 14.3% in the diuretic-based regimen will develop overt diabetes mellitus 21 . The detectable risk increase is compatible with 80% power and significance of 0.05 is 49% (OR 1,489). Based on this assumption, a sample size of n = 2001 was required. In terms of randomisation, addressing the literature evidence in favour of ACE inhibitor treatment, an imbalanced ratio of 2:1 for the number of patients in both treatment groups was chosen, resulting in a target inclusion number of 1334 patients in group 1 and 667 patients in group 2, respectively.

The following parameters will be analysed: patient demographics including patient history, capillary und venous fasting glucose, HbA1c, BP, BMI, waist/hip ratio, concomitant medications, percentage of patients with prediabetes according to PreDiSc parameters 22 , total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, urinary albumin, and high-sensitive C-reactive protein.

Patients treated with ramipril (and ramipril-based combination therapy) and patients who received various other antihypertensive drugs (with the exception of ACEi or ARBs) will be compared. Statistical analyses of the data will be performed as exploratory analyses. Descriptive statistics for continuous target data per treatment group and per total contain the following: number of patients, means ± standard deviation, median, minimum and maximum. The absolute and relative frequency in percentages will be determined. Per treatment group 95% confidence intervals for the means of continuous target data as well as for the relative frequency of categorical target data will be calculated using appropriate methods.

For comparison of the treatment groups with respect to the incidence of specific events (e.g. patients with first manifestation of diabetes mellitus type 2 or proportion of patients with deterioration of pre-diabetes), the chi-square or the log-rank tests will be used to compare the "survival curves". An interim analysis is scheduled after the first and second year of the observation. However, the overall significance levels will not be adjusted.

The primary goal of AdaPT is to compare the effects of two antihypertensive combination therapies, an ACE inhibitor based treatment with a diuretic- (or betablocker)-based treatment on the incidence of new-onset of type 2 diabetes. To provide adequate power to discriminate a potential differential effect of these therapies, the trial is being conducted in pre-diabetic patients with hypertension and metabolic disorders in which there is a high probability for the development of diabetes.

AdaPT is conducted as an observational study in daily practice in Germany allowing for the widest possible applicability of the results obtained. This is important because there is considerable heterogeneity in patient management in daily practice and patients with pre-diabetes in clinical trials usually differ substantially from those in clinical trials. Taking the PreDiSc Score as a screening algorithm, the study allows physicians to effectively screen for a high risk for the development of diabetes. This would, in case of positive results, serve as an easy screening tool for high risk patients in the future.

Compared to the previously mentioned trials AdaPT included patients at a higher risk for the development of diabetes based on the selection criteria age, presence of hypertension, an impaired fasting glucose and the missing exclusion of prior cardiovascular disease. As opposed to DREAM (Placebo control) AdaPT includes patients treated with betablockers or diuretics as a control group. This further enhances the likelihood of a differential effect based on the previously mentioned data by Abuissa 32 , Elliott 15 and Lam 14 (see also Figure 1). Together with the 4 year follow-up it appears likely that AdaPT may document a reduced incidence of diabetes.

The fact that a score (PreDiSc) is applied in the ADaPT investigation to prospectively identify eligible patients represents a novel approach, since it shifts the conventional focus from individual risk factors (hypertension, dyslipidaemia) to a more comprehensive view that considers absolute patient risk 3 33 . To our knowledge, such an approach has only been pursued in retrospective post-hoc investigations, for example in a current analysis of the LIFE study 34 .

The oral glucose tolerance test (OGTT) is the standard screening test in high risk populations (identified by medical history), but the fasting plasma glucose test is more convenient under daily practice conditions. 35 Determining fasting plasma glucose lacks sensitivity however and may miss a number of patients with diabetes. Another possible variable to determine glycaemic control is HbA1c but is less suitable for a general screening. 36 37

Against this background, the Pre-Diabetes Score (PreDiSc) Study established a set of easy-to-determine clinical and/or laboratory parameters with close correlation to the outcomes of an OGTT 22 . Indeed, using the HbA1c alone yielded low sensitivity (58%) as did fasting glucose alone (62%). However, a combination of age ≥ 55 years, systolic BP of ≥ 140 mmHg, fasting glucose ≥ 110 – 126 mg/dl, elevated HbA1c ≥ 6% and abdominal obesity (waist circumference > 88 cm in women and 102 cm in men) was associated with high sensitivity of IGT (i.e. identification of individuals with pre-diabetes: 79%) as well as high specificity (i.e. exclusion of individuals without pre-diabetes: 74%) 22 . The low acceptance of the OGTT and the non satisfying sensitivity and/or specificity of the sole HbA1c determination speak against these parameters as inclusion criteria for studies in daily practice. In contrast, the satisfactory predictive value of the clinically easy to determine dataset evaluated in PreDiSc was the rationale to use this score as screening procedure in a prospective study. Hypertensive patients fulfilling the PreDiSc criteria have a very high likelihood to progress to overt diabetes. Compared to the original PreDiSc score, on the basis of practical experiences in the initiation phase of the study two amendments to the ADaPT protocol became necessary: first, HbA_1c was to be measured in a central laboratory instead of the originally foreseen local laboratories (due to wide variation in locally determined values). Second, the age criterion was reduced to 45 years or older in order to accelerate the inclusion rate.

Certain limitations of the investigation have to be taken into account: first, it is controlled, but not randomised. This means that the comparability of both cohorts can be assessed retrospectively in terms of known factors that may influence the outcomes (e.g. age, gender, comorbidity), but not in terms of unknown bias. Second, the majority of patients will require combination therapy of two, three or even more antihypertensive drugs of various classes. At clinical practice conditions, a substantial proportion of patients may receive "unallowed" medications in the course of the investigation (i.e., from the regimen of the other arm). On the other hand, the real practice conditions in this investigation convey substantial benefits that will extend the knowledge from randomised studies. Patients will be less selected than typical study patients. The dosing regimens and the combinations will reflect current use (e.g. lower doses as compared to the US for some drugs, inclusion of ARBs, etc.) and thus the results can be easily extrapolated to day-to-day clinical use.