Non-invasive hemodynamic monitoring as a guide to drug treatment of uncontrolled hypertensive patients: effects on home blood pressure in the BEAUTY study

Abstract

Background: In the BEAUTY study we investigated whether utilizing non-invasive monitoring of hemodynamic parameters combined with a drug selection algorithm (integrated hemodynamic management-IHM) compared to conventional drug selection may improve home BP in patients with uncontrolled hypertension.

Methods: Uncontrolled (office systolic blood pressure (SBP) > 140 mmHg and ambulatory daytime SBP >135 mmHg while taking ≥2 antihypertensive drugs) essential hypertensive patients were referred to 5 European Hypertension Excellence Centers and, if eligible, were randomized into IHM-guided vs conventional treatment adjustment. Home blood pressure (BP) was taken with 2 repeated readings at 1–2 min intervals in the morning and in the evening (before drug intake and eating) during the week preceding the visit at the outpatient clinic after 5 min rest using a validated semi-automatic oscillometric arm cuff device and with a correct cuff bladder placement. Home blood pressure was measured in a sub-group of patients (n = 84) not significantly different from the other patients.

Results: Home SBP changed from 152.1+/−15.8 and 149.8+/−11.8 mmHg to 131.0 +/−11.1 and 139.6+/−12.8 mmHg in IHM group (n = 46) and Control group (n = 38), respectively, showing significantly greater reduction in IHM than in Control group (d= −10.9 mmHg, 95% CI −17.77, −4.02), p = 0.002. The reduction remained significant after multiple adjustments, particularly for baseline home SBP, recruiting center, age, sex and BMI (SBP_IHM-Control= −9,63 mmHg, 95% CI −14.28, −5.11) mmHg, p < 0.0001).

Conclusion: Drug selection algorithm based on non-invasive hemodynamic monitoring induced larger reduction in home BP compared to conventional drug selection in uncontrolled hypertensive patients referred to European Hypertension Excellence Centers. Although the main BEAUTY study was negative, these home BP measurements taken by patients themselves may suggest that the integrated hemodynamic monitoring is useful in patients with uncontrolled hypertension. This finding might depend on specific features of home BP measurements which could make it recommended BP measurement method for drug trials.

Keywords:

• Antihypertensive drugs
• blood pressure control
• integrated hemodynamic monitoring
• uncontrolled hypertension
• home blood pressure monitoring

Introduction

Appropriate control of BP is essential for prevention of future cardiovascular events. However, BP control among treated hypertensive patients is generally considered as insufficient. There is growing evidence linking hypertension to altered hemodynamic modulators [1]. Better selection of antihypertensive drugs, guided by non-invasive integrated hemodynamic management (IHM) using impedance cardiography, may possibly counteract such problems and provide better BP control [2,3]. In the BEAUTY study [4] non-invasive hemodynamic monitoring associated with a drug selection algorithm induced similar reductions in ambulatory daytime and office SBP compared to conventional drug selection.

Recently, the usefulness of self-measured BP at home (home BP measurement) for the management of hypertension has been reported in many studies. When compared with office BP, home BP yields multiple measurements over several days taken in the individual’s usual environment and is more widely available [5]. A large systematic review has demonstrated that home BP is a significant predictor of cardiovascular mortality and cardiovascular events and an important prognostic variable over and above that of office BP [6].

Thus, the aim of the present study was to explore whether IHM improves the blood pressure control according to home BP measurements and test the hypothesis that difference may be visualized better with home BPs.

Materials and methods

Study design, patients and objectives

The BEtter control of blood pressure in hypertensive pAtients monitored Using the hoTman^® sYstem (BEAUTY) Study was an investigator initiated and driven multicenter prospective randomized parallel groups controlled study with sites at five Excellence Centers of the European Society of Hypertension (ESH), Gdansk (Poland), Milan (Italy), Oslo (Norway), Paris (France), and Tallinn (Estonia), ClinicalTrials.gov Identifier: NCT01482364. The study was overseen by a steering committee consisting of the five senior investigators representing all sites (SEK, MV, KN, GP, SL).

Primary objective was to explore whether monitoring hemodynamic parameters and then applying a predefined algorithm of drug selection (i.e. IHM) with the HOTMAN^® System improves the control of daytime SBP at ambulatory blood pressure monitoring (ABPM) in hypertensive patients, as compared to classical drug selection (i.e. without IHM) during a 6 month intensive treatment program. One of the secondary objectives was to explore whether IHM improves the blood pressure control according to home BP measurements, which is the focus of the current paper.

Patients aged 18–85 years of either gender were referred from general practices responding to letters of invitation or were recruited directly by newspaper advertisements or referrals to the university outpatient clinics. They were worked-up at the five participating university hospitals in the time period from January 2011 through December 2012 by experienced physicians. Of the 315 screened patients 167 met the inclusion criteria and were randomized [6] and outcome was available for 156 patients (full analysis set). Among these 156 patients, 84 (54%) entered into the substudy that included also home blood pressure measure (HBPM). This subgroup was not significantly different from the subgroup without HBPM. Oslo center did not do home BP as it was not mandatory in the main protocol.

UCH was defined as elevated office SBP (>140 mmHg) despite regular intake of 2–4 or more antihypertensive drugs (towards the end of recruitment we accepted up to 7 drugs but in few patients only, which is reflected by a total average of 3 drugs). Additionally, patients had to qualify by also having mean ambulatory daytime SBP >135 mmHg. Drug treatment was unchanged for two weeks and no other change in medication was pre-planned for the following 6 months.

The study was approved by all relevant committees for clinical research ethics in the participating countries and by the institutional research committee of all 5 hospitals. All patients gave written informed consent for participation and publication of results. All patients who qualified for the procedure within the 24 month time period were included. Expenses were mostly covered by the hospitals and partly by grants-in-aid from one sponsor (Hemo Sapiens Inc. European Office, Bucharest, Romania), and patients were not paid.

Patients were randomized using a pre-determined 2 × 2 randomization list through a website organized by the monitor of the study (Sintesi Research, Milan, Italy).

Home BP measurements

Home BP was available in 46 IHM and 38 controls at the beginning of the study and at 6 months. Home BP was measured for one week before each study visit. Home BPM was done according to the 2008 European Society of Hypertension (ESH) home BP monitoring guidelines [7]: a) with 2 repeated readings at 1–2 min intervals in the morning and in the evening (before drug intake and eating); b) during the week preceding the visit at the outpatient clinic; c) after 5 min rest and 30 min without smoking or caffeine, seated, back supported, immobile, legs uncrossed, not talking, relaxing, arm resting on the table; d) using a validated semi-automatic oscillometric arm cuff device and with a correct cuff bladder placement; e) a written report of BP values was provided to the physician during the outpatient clinic visit; and f) the first day of each weekly session was discarded. Home SBP and DBP (diastolic blood pressure) were determined as the average of all remaining morning and evening values for the period considered. The effects of each management strategy were assessed by comparing the home BP values obtained in two occasions, i.e. before the randomization visit and before the last visit, scheduled.

Validated home BPM device based on the same software, such as the Microlife Watch-BP device, which also offered the possibility to perform and store home BP measurements following the 2008 ESH home BPM guidelines. Home BP measurements were not performed on all patients in the study due to shortage of devices in some centers and due to home BP being a secondary endpoint and not obligatory in the protocol.

Impedance cardiography with the HOTMAN^® (hemodynamic & oxygen transport management) system

The HOTMAN^® system allows a non-invasive assessment of patient’s hemodynamics with two important improvements compared with other previous similar devices: (i) use of a very low current (7 μA, 300–400-fold lower than that used by other products, making it safer for the patient), and (ii) use of a new data signal processing and of an improved mathematical algorithm. The system has been validated against an invasive thermodilution approach [3].

Recordings by the HOTMAN^® System are performed with patients in the supine position, resting for 5 minutes before measurement and involve non-invasive measurement of thoracic impedance through placement of four pairs of thoracic electrical bio-impedance specific sensors placed on the neck and lower thorax. Electrical impedance changes are digitally processed to calculate different hemodynamic parameters. Compared to classical ‘impedance cardiography’ the HOTMAN^® system has some different characteristics. It measures SVRI per beat, i.e. SSVRI (Stroke systemic vascular resistance index), an index of vasoactivity, which is not detected by classical systems.

Adjustments of drug treatment which could be done by changing either class of medication or their doses, were then performed in the IHM group under the guidance of a pre-specified algorithm according to patients vascular resistance, volemia and/or inotropy [7] (Table 1).

Table 1. Scheme for selecting and titrating antihypertensive drugs.

	Antihypertensive drug
Vasoconstriction	If +34 to +100%, then use normal dose of CCB-DHP, ACEI, ARB		If > +100%, then use high doses of CCB-DHP, ACEI, ARB
What to do	In each case:-if inotropy > +20%, then use CCB-DHP-if inotropy is normal (−20% to +20%) or low (i.e. hypo-inotropy, less than –20%), then use ACEI or ARB
Hypervolemia	If normal (−20 to +20%), then no DIU		If high (>+20%), then use DIU
What to do			If +20 to +50%, then use normal dose of DIU	If > +50%, then use high dose of DIU
Hyperinotropy	With normal volemia		With DIU-induced hypovolemia (<–20%)
What to do	If +20% to +60%, thenuse VD-BB	If > +60%, then use non VD-BB	Stop DIU,Add CCB, ACEI or ARB Do NOT prescribe BB
Hypoinotropy	Stop BBPrefer CCB-V or CCB-D rather than CCB-DHP

Conversely the control group had their antihypertensive medication adjusted at baseline, 1 month and at 3 months by lead investigators in each site according to 2013 ESH/ESC hypertension guidelines [8] only. This includes utilizing home BP data whenever available in approximately half of the patients but investigators were free to choose drugs according to their expertise and specific instructions on choice of drugs were not given in order avoid confounding the trial by an ‘intervention in the intervention’ bias [9].

Monitoring and data handling

The study was monitored by Sintesi Research (Milan, Italy), an independent company with no relationship to the sponsor. Sintesi Research monitored each site according to the GCP standard, and cleaned queries prior to transferring data to the central database in Milan, and before locking of the database prior to any statistical analysis.

All impedance cardiography measurements and ambulatory BPs were stored together with all the assembled data in a centralized database at Instituto Auxologico Italiano, Department of Cardiology, Milan, Italy. All statistical work included in this paper was done by professional independent medical statisticians (PR, XL, MV) in University of Milano-Bicocca.

Statistical analysis

Descriptive statistics, including means, standard deviations, and frequencies, were used to summarize the patients characteristics. Characteristics of patients with and without HBPM were compared by Chi-square test and t-test for categorical and continuous variable, respectively. Fisher’s exact test was used when there was at least one cell with expected count less than 5.

The difference among the two arms on HBPM, ABPM and OBPM were evaluated using two-sample t-test on BP changes from baseline to follow-up. Chi square test with one degree of freedom was used to compare the percentage of normalization of BP.

A linear regression model on HBPM and ABPM at follow-up (M6), adjusting for the baseline home SBP, was applied to account for centers and potential risk factors (sex, age, BMI) in the estimate of treatment effect. Rate of fall (slope with time) in OBPM was analyzed separately by a longitudinal analysis. A mixed model was applied including (as response variable) all the BP measurements available from baseline to M6 for each patient on the secondary analysis set (with at least two measurement available and assuming missing at random). The fit of a linear (decreasing) trend by visit was checked by a spaghetti plot. The dependence between measurements on the same subject was accounted for by the inclusion of a random intercept in the model. The model included as regressors the visit number/time of the visit, the randomisation arm, the interaction between the arm and time and other potential risk factors.

All tests were bilateral using α = 0.05. Data were recorded in a SAS database and analyzed using SAS 9.2.

Results

Patients

The characteristics at baseline of 84 uncontrolled hypertension (UCH) patients randomized to IHM adjusted drug treatment (n = 46) and to classical clinical adjustment of medical treatment (n = 38) are compared in Table 2. There was no difference regarding demographic characteristics, disease history and number of antihypertensive drugs between groups. Patients used tolerated doses of an average of 3.2 antihypertensive agents (ranging from 2 to 6). Mean number of agents per patient has increased during the study by 0.7 ± 1.09 in the IHM group and 0.5 ± 0.95 in the Control group.

Table 2. Demographics of subjects who had home BP at both V2 and V6.

HOME BP measurement:		Yes			No	P-value of χ^2 comparing pts with and without HOME BP
		Total N pts	IHM group	Control group	Total N pts
			N (Percentage, %)	N (Percentage, %)
All		84	46(100)	38(100)	72
No. of subjects by Center	France	1		1(3)	2	<0.001^Table Footnotea
	Italy	10	6(13)	4(11)	9
	Poland	24	12(26)	12(31)	10
	Estonia	49	28(61)	21(55)	1
	Norway	0			50
Sex	Male	45	24(52)	21(55)	48	0.0966
	Female	39	22(48)	17(45)	24
BMI continuous, kg/m2	Mean (standard deviation)	30(4)	30(4)	30(4)	29(4)	0.1632^b
BMI categorical	≤24.9	13	6(13)	7(18)	14	0.1661
	25–29.9	28	20(44)	8(21)	32
	≥30	43	20(44)	23(61)	26
Race	Caucasian	84	46(100)	38(100)	69	a 0.0961
	Black	0			3
Smoking habits	0 (non-smoker)	78	44(96)	34(90)	62	0.3243
	1–3	1	1(2)	–	3
	≥4	5	1(2)	4(11)	7
Alcohol habits	0 (teetotal)	8	4(9)	4(11)	9	0.3469a
	occasional	69	38(83)	31(82)	52
	1 drink/day	3	2(4)	1(3)	7
	2 drinks/day	4	2(4)	2(5)	3
	>2 drinks/day	0			1
No. of antihypertensive agents at baseline (Visit 2)	Mean (standard deviation)	3.2(1.1)	3.0(1.0)	3.4(1.1)	3.2(1.0)	0.7491b

^aFisher exact test.

^bt-test.

Note: data are expressed as number (percentage) for categorical variables and mean (standard deviation) for continuous variables.

Office, ambulatory and home blood pressure changes between study end and baseline

In the subgroup of patients with HBPM, mean office SBP changed from 159 ± 23 and 155 ± 16 at baseline to 140 ± 17 and 140 ± 16 mmHg at study end in IHM group and Control group, respectively (d= −4.08 mmHg, 95% CI −12.81;4.65, p = 0.36). In the same subgroup, mean day-time SBP by ABPM changed from 148 ± 12 and 149 ± 10 mmHg at baseline to 133 ± 12.0 and 133 ± 12 mmHg at study end in IHM and Control group, respectively, with no difference (d) between the two groups (d= −0.58 mmHg 95% CI −6.11;7.28, p = 0.86). The linear regression model adjusting for baseline SBP, recruiting center, age, sex and body mass index (BMI) confirmed no difference between the two groups in ambulatory SBP at 6 months (SBP_IHM-Control = −0.93 mmHg, 95% CI −5.9; 4.05 mmHg, p = 0.72).

Home blood pressure changes between study end and baseline and control rates at study end, are demonstrated in Table 3.

Table 3. Home blood pressure values and control rates at 6 months.

Changes in HMBP (visit 6-visit 2)	IHM group (N = 46)	Control group (N = 38)	Diff (IHM-Control)	P-value
	Mean (SD)	Mean (SD)	Mean (95% CI)
Home SBP	−21.1 (17.7)	−10.2 (13.0)	−10.90 (−17.77, −4.02)	0.002
Home DBP	−7.6 (9.0)	−6.9 (10.6)	−0.67 (−4.91, 3.58)	0.756
SBP and DBP normalization at 6 months	n (%)	n (%)
Home SBP(<135 mm Hg)	30 (65%)	12 (32%)		0.002
Home DBP(<80 mmHg)	34 (74%)	27 (71%)		0.770

Home BP displayed at 6 months a significantly greater reduction in IHM (−21.1 ± 17.7 mmHg) than in controls (−10.2 ± 13.0 mmHg, P = 0.002). Home SBP changed from 152.1 ± 15.8 and 149.8 ± 11.8 mmHg to 131.0 ± 11.1 and 139.6 ± 12.8 mm Hg in IHM group and Control group, respectively, showing significantly greater reduction in IHM than in Control group (d= −10.9 mmHg, 95% CI −17.77, −4.02, p = 0.002, Table 3), which remained significant after multiple adjustment particularly for baseline home SBP, recruiting center, age, sex and BMI (SBP IHM-Control= −9.63 mmHg, 95% CI −14.28, −5.11, p < 0.0001). Figure 1 demonstrates results of ABPM and HBPM at the beginning and at the end of the study in the IHM group compared to the Control group.

Figure 1. Comparison of home and ambulatory BP in the two groups.

The key finding of the present study is a statistically significant difference in the decrease in home BP between groups.

Discussion

The main finding of the present study is that home BP displayed at 6 months had a significantly greater reduction in IHM than in controls, whereas office did not. This emphasizes the superiority of HBPM over office BP measurements in detecting different BP effects induced by different management strategies.

The primary function of our cardiovascular system (CV) system is not a generation of BP but a delivery of oxygen to all tissues. BP is a measurable end product of an exceedingly complex series of factors, including those that control blood vessel caliber and responsiveness, those, which control fluid volume within and outside the vascular bed, and those, which control cardiac output. Data suggest that there is a strong relation between hypertension and abnormal hemodynamic modulators. In our previous paper we have demonstrated, that almost all (98.5%) uncontrolled hypertensive patients in the ESH Excellence centers presented at least one altered hemodynamic modulator: intravascular hypervolemia (96.4%) and/or hypoinotropy (42.5%) and/or vasoconstriction (49.3%) [1].

Because BP elevation is commonly treated like a symptom, without paying attention to the hemodynamic causes, physicians often neglect the other hemodynamic parameters like cardiac output, left ventricle contractility, and vascular resistance. It is all the more surprising since antihypertensive drugs modify the entire hemodynamic status. This could be an explanation for the relatively low rate of BP control in the hypertensive population.

In the current analysis of the BEAUTY study, we focused on home BP measurement in response to either drug selection algorithm according to non-invasive hemodynamic monitoring or conventional use of drugs in the Control group. The advantages of home BP monitoring in the management of treated hypertensive patients have been reported in many studies [16,17]. During the last years, guidelines have placed much greater emphasis on the utilization of out-of-office measurements including home BP measurements [8,10,11]. Home BP measurement has been demonstrated to be concordant with ambulatory blood pressure monitoring. Paolasso et al. reported no significant differences in the measurement of diurnal SBP and DBP between both methods [12]. Home BP values are obtained under stable conditions and can eliminate the white coat effect [13]. Use of home BP monitoring according to ESH guidelines, i.e. by considering the average of duplicate readings taken over a week before each office visit, after discarding the values measured during the first day (Parati et al. J Hypertens. 2008) yields highly reproducible values which are particularly appropriate for the management of hypertensive patients receiving antihypertensive drugs [13].

In the present study we have used validated home BP devices based on the same software as the devices used for 24h ABPM, i.e. the Microlife Watch-BP device, which also offered the possibility to perform and store home BP measurements following the 2008 ESH home BPM guidelines.

Our findings are in line with the evidence supporting the use of out-of-office monitoring in all aspects of routine clinical care, which has increased substantially in recent years and is reflected in an increased utilization of home BP monitoring by patients and clinicians [14]. Home BP self-measurement and monitoring improves patient’s awareness and their adherence to prescribed treatment, thus favoring a better management of hypertension. This represents an important complementary support to the doctor-patient relationship in the office, with a high potential for improving hypertension management [15,16].

In our previous paper we have found no statistically significant difference between the decrease in office BP in IHM group compared to Control group [6]. We did not find either a significant difference in BP changes between groups when using ABPM, and these findings were replicated in the present subanalysis. In contrast, the key finding of the present study is the occurrence of a statistically significant difference in the decrease in home BP between groups. The reason for these discrepancies is not clear. We hypothesize that this finding might depend on specific features of home BP measurements which could make it best BP measurement method for drug trials or intervention studies. Home BP monitoring provides more reproducible data on an individual’s BP; obtained at rest, repeated daily, and eliminating white coat effect. Moreover, home BP is more similar to «basal» blood pressure. The conclusion we can draw from our data is that home BP is more sensitive to drug selection algorithm in response to non-invasive hemodynamic monitoring than office and ABPM in patients with uncontrolled hypertension.

Study limitations

The present study must be interpreted within the context of the potential limitations. First, home BP measurements were not done in all BEAUTY study patients due to shortage of devices in some centers. Second, home BP monitoring was performed in 4 study centers out of five, home BP being a secondary endpoint and thus not obligatory according to the study protocol. However, in spite of these limitations, we have found a statistically significant greater reduction in home systolic blood pressure using IHM than in the control group. Further research is required in larger cohorts of patients, but our data clearly support the importance of out-of-office BP monitoring is in hypertension management.

Conclusions

In conclusion, easy-to-do non-invasive hemodynamic monitoring associated with a drug selection algorithm induced larger reduction in home BP compared to conventional drug selection in uncontrolled hypertensive patients referred to ESH Excellence Centers. Home BP monitoring appears to be a convenient method for BP measurement in clinical trials. These exploratory data should stimulate further studies aimed at investigating IHM-guided hypertension management in a larger number of subjects followed up not in excellence centers for hypertension but rather in in daily practice.

Acknowledgements

The research was funded partly by the Estonian Ministry of Education and Research under institutional research financing IUT 19-2. Paola Rebora was supported by the grant SIR RBSI14LOVD of the Italian Ministry of Education, University and Research.

Disclosure statement

S.E.K. has received lecture honoraria from AZ, Bayer, Medtronic, MSD, Novartis and Takeda, honoraria for consulting from Bayer, Medtronic, Takeda and Serodus, and research support from AZ and Pronova. G.P. has received lecture honorarium from Hemo Sapiens, Daichii Sankyo, Omron, and Servier. S.L. has received lecture honoraria from Daichii-Sankyo, Novartis, Omron, and Servier, and research support from Atcor, Esaote Pie Medical, and Servier. K.N. has received lecture honoraria or from Servier, Krka, Berin-Chemie/Menarini, Egis, Sandoz, Mylan, Polpharma, Adamed and Gedeon Richter. M.V. has received lecture honoraria from AZ, Boehringer Ingelheim, Servier, Abbott and Menarini. The other authors have no disclosures.