We studied nine male patients, mean age 74 years, (range 58-79 years), with severe coronary artery disease (eight with three-vessel and one with two-vessel disease) and severely symptomatic (CCS angina functional class II-III). Except for one patient who had dynpnea, the rest suffered from daily angina. Eight of the patients underwent later coronary bypass grafting (CABG) and one patient with earlier CABG underwent percutanous coronary intervention. All patients had normal global LV systolic function at rest except for one patient who had mildly depressed global LV systolic function (LV ejection fraction ? 45%). Three patients had a history of myocardial infarction, three had a history of systemic arterial hypertension, three had a history of paroxysmal atrial fibrillation, three had a mild form of chronic renal failure, and none had a history of diabetes mellitus. All patients were treated with aspirin, beta-blockers, long acting nitrates and calcium channel blockers and seven with either angiotensine converting enzyme inhibitors or angiotensine II receptors blockers.

We obtained a baseline two-dimensional echocardiography with superimposed Tissue Doppler Echocardiography (TDE) images using a multi-Hertz transducer with commercially available equipment (Vivid 7, GE Vingmed, Horten, Norway) for all patients prior to bicycle exercise stress echocardiography test. Standard parasternal short and long axis views, as well as apical two, three, and four chamber views, were digitized during three consecutive cardiac cycles in cine loop format for off-line analysis.

The patients underwent two separate exercise stress echocardiography test in semi-supine position on a specially designed, table-mounted bicycle ergometer. The initial workload was 25 Watts with increments of 10 Watts added every minute. At each stage, the heart rate, systolic blood pressure, and a 12-lead ECG were recorded. We obtained standard echocardiographic images of five consecutive cardiac cycles containing TDE information at rest, sub-maximal workload (heart rate = 70% of the expected maximal heart rate), maximal workload, and recovery phases (one to two minutes after the exercise); all measurements were digitally recorded and stored. Stopping criteria for exercise test was either severe chest pain or achieved 85% of maximal expected heart rate (220-age).

An intravenous infusion of either low-dose adenosine (35 ?g/kg/min) or placebo (saline infusion) was given in a double-blind fashion over 10 minutes prior to the start of the exercise test. The order of the drug administration was randomly selected and the sequence of infusions was inverted on the second occasion; the first and second occasions occurred at least one day apart, median 3 days (range 1 to 5 days) (Figure 1). The rationale to select this dosage lies on previous studies showing that it had an analgetic effect[⁸], doses between 50 and 70 ?g/kg/min may have a positive effect while doses higher than 70 ?g/kg/min provoked ischemia[⁹] in patients with coronary artery disease.

The analysis of the echocardiographic images was done by one of the investigators (MQ) blinded to the sequence of drug administration. However, the images corresponding to each stage of the protocol were known as the heart rate at rest and at peak exercise were clearly discernible. A sample volume was positioned on the region of interest at the basal segment of the four LV walls (septal, lateral, inferior, and anterior) to obtain a myocardial velocity profile (Figure 2). We analyzed both the systolic and diastolic phases of the velocity profile. We defined peak systolic velocity (PSV) as the peak velocity during ejection. We also measured peak velocity at early diastole (E'-wave) and peak velocity at the late diastole (A'-wave). The longitudinal A-V plane displacement was calculated by time integration of the PSV. In addition, the degree of myocardial compression/deformation as strain (S) and strain rate (SR) were also calculated.

In order to assess the LV regional systolic function and to detect the presence of regional myocardial ischemia, the LV walls were categorized as ischemic or non-ischemic: a LV wall was defined as ischemic if a reduction, no increment, or an increment of <15% in PSV, was observed during maximal work load compared to baseline preceding placebo-infusion; otherwise, the LV walls were defined as non-ischemic.

Data are presented as mean ? standard deviation (SD). Statistical comparisons were performed by analysis of variance (ANOVA) with repeated measures. Post hoc comparisons between groups were performed using the Tukey test.

The local Ethics Committee at the Karolinska University Hospital, Huddinge, approved the study. All participants gave written informed consent.

The small size of this study precludes the extrapolation of our results to all patients with ischemic heart disease, especially to those with mild forms of exercise-induced ischemia. The positive pre-conditioning effects of adenosine were not observed for all parameters reflecting the regional LV systolic function. In fact, strain rate, a measure of "pure" myocardial contractility, was not affected by adenosine. This result could be explained by the low reproducibility of this measure when assessed during exercise[²⁰].

Certainly, some degree of ischemic preconditioning induced by angina pectoris did exist among all the patients. However, we do not believe that this fact compromised the results of the study, considering the placebo/controlled and the cross-over character of the study. We used high ischemic limit criteria to define myocardial ischemia by tissue Doppler echocardiography, and when milder criteria that allow for the detection of less severe forms of ischemia were applied, the results were the same. However, if we applied the same ischemic criteria as are used in clinical practice[²¹], almost all LV walls were categorized as ischemic, making it impossible to perform adequate statistical analysis. This fact emphasizes that we were studying patients with severe coronary artery disease: severely symptomatic and with severe exercise induced myocardial ischemia.

The present results may open the possibilities of testing new therapeutic agents for the treatment of not only high risk patients with preconditioning mimetics that stimulate the biochemistry of preconditioning without actually causing ischemia but also of a vast number of patients with stable angina pectoris with mild to severe symptom in whom coronary revascularisation is impossible or not achievable.

Values are expressed as means ? SD.

Abbreviations: A' = late diastolic A-wave; A-V = atrio-ventricular plane displacement; bpm = beats per minute; E' = early diastolic E-wave; PSV = peak systolic velocity; SBP = systolic blood pressure, S% = strain; SR = strain rate.

* P < 0.01 Peak versus baseline

Values are expressed as means ? SD.

As in table 1

* P < 0.01 Peak versus baseline

? P < 0.01 Peak - ischemic walls - adenosine, versus peak - ischemic walls - placebo