Ambulatory Blood Pressure Monitoring in Patients With Coronary Artery Disease

Study design and patient population We included 225 patients scheduled for follow up after an acute or elective percutaneous coronary intervention (PCI) at the Department of Cardiology at the Karolinska University Hospital 2009-2015. Exclusion criteria were age \<18 or >90 years. Information on medical history, date and indication for PCI were collected from the medical records, by the study physician (OH).

All patients were subject to ABPM at baseline (3-6 weeks after PCI) and follow up (11-18 weeks after PCI. At the time of the first ABPM the patients also completed a questionnaire about smoking habits and current antihypertensive drug treatment. Weight (kg) and height (cm) was measured. At the clinical follow up visit one to two weeks later they were assessed by a cardiologist not involved in the study. At this visit ABPM results were either used or not according to randomization in the decision making for adjustments in antihypertensive medication. The cardiologists were provided with reference values for ABPM according to guidelines.26,27 For those with concealed ABPM-results the OBP was used for decision making regarding changes in antihypertensive treatment. Finally, results regarding BP control were assessed at the second ABPM 8-12 weeks following the first measurement.

Office Blood pressure measurement OBP was recorded in both arms by a biomedical scientist or a specialized nurse (study staff) using a mercury sphygmomanometer with the subject in the supine position after 5 min of rest. The mean of two consecutive readings was calculated. If there was a difference in systolic or diastolic BP (SBP, DBP) between the arms of >10 mmHg, the arm with the highest reading was used when defining OBP, otherwise the non-dominant arm was used. The same arm was used at the follow up where either the physician or a nurse measured OBP after having been given instructions for standardized BP measurement as described above.

Ambulatory BP Ambulatory BP values were obtained using a noninvasive oscillometric system (Spacelabs 90217, Spacelabs Healthcare, Hertford, UK). The device was fitted to the patient by one of the study staff. Patients were instructed not to restrict their daily activities during the monitoring periods. Before the start of the monitoring period, the automatic readings were cross-checked against manually measured BP by auscultation. The device was fitted to the nondominant arm, unless a difference of >10 mmHg in SBP between the arms was recorded, in which case the arm with the highest pressure was used. BP and heart rate were recorded automatically every 20 minutes' daytime and every hour at night for a 24-h period. The BP data was auto-edited by the Spacelabs program, which excluded presumably erroneous data. No manual editing of data was carried out in order not to induce bias. Means were calculated for the whole 24-h period, and for day (07.00-21.00) and night (24.00-06.00) periods separately.

Antihypertensive drug treatment The agents according to guidelines26 27 classified as BP lowering were thiazide- and potassium saving diuretics, beta-blockers, calcium antagonists, ACE inhibitors, angiotensin II receptor blockers and others (doxazocin only one used). At the follow up antihypertensive treatment changes were reported by the physician in the study protocol. The patients' current antihypertensive treatment was also documented prior to the 2nd ABPM by the study staff. All antihypertensive treatment was further recorded as a percentage of recommended maximal daily doses, to allow for calculation of treatment change. The latter was calculated as the difference in antihypertensive treatment between prior to the physician follow up and ongoing medication at the 2nd ABPM.

Statistical analysis Mean and standard deviation (SD) were used for numerical data whereas median and range was used for the number of days until follow-up visit and the number of BP-lowering agents. Chi-square-tests or Fishers exact test were used to compare ratios between groups where the variables were nominal. Independent t-tests were used to compare continuous variables between groups, since the variables investigated were normally distributed. In order to study whether the intervention optimized BP across the distribution of baseline ABP, a multivariate analysis was performed and the interactions between study groups on the association between baseline ABP and change in ABP as well as change in antihypertensive therapy was calculated. Separate analysis was performed in the diabetic subgroup. A p-value \<0.05 were used to define statistical significance. Statistical analysis was done in the StatSoft program STATISTICA®.

Ethical considerations Ethical approval was applied for and approved by the Stockholm Regional Ethical Review Board, reference number 2008/1017-31. All subjects gave informed consent.