Is BP measurement at home or in the physician%u2019s office more reliable?

Previous research has shown that modifying blood pressure treatment according to ambulatory measurements rather than using office measurements has resulted in a decrease in treatment and a reduction of left ventricular hypertrophy, while maintaining control of hypertension and general well-being.1 However, this approach did not lessen the overall costs of antihypertensive treatment.

The benefits of ambulatory monitoring, such as the lack of white-coat syndrome, absence of observer bias when automatic instruments are used, and larger number of readings, are also present with self-measurement of blood pressure by the patient at home. In addition, self-monitoring costs less than ambulatory monitoring and may increase adherence to antihypertensive therapy.

We designed the Treatment of Hypertension Based on Home or Office Blood Pressure (THOP) trial to examine the effect on adjustment of antihypertensive medications resulting from home blood pressure (HBP) measurement compared with office blood pressure (OBP) monitoring.

Patients and methods

The protocol of the THOP trial has been described in detail previously.2 After a 1-month run-in period, hypertensive patients (sitting diastolic blood pressure > 95 mm Hg) were randomly assigned to receive treatment based either on self-measured diastolic HBP or on diastolic OBP. Regardless of randomization, the blood pressures for both the HBP group and the OBP group were measured at each office visit. HBP was the mean of 42 readings, that is, three readings in the sitting position in the morning and three readings in the evening over 7 consecutive days, taken with an oscillometric Omron HEM-705CP device.3 OBP was the average of three readings in the sitting position taken by the physician at the office with a sphygmomanometer. At baseline and after 6 and 12 months, daytime (from 10:00 am to 8:00 pm) ambulatory blood pressure (ABP) was calculated from recordings obtained with oscillometric SpaceLabs 90207 recorders.3 The ambulatory measurements were not considered in any treatment decisions.

In both groups, treatment was adjusted in a stepwise manner to reach the target range of diastolic blood pressure, which was 80 to 89 mm Hg. Once patients were randomly assigned to groups, they were started on monotherapy with 10 mg of lisinopril (Prinivil, Zestril) per day (step I). At each follow-up visit, treatment could be increased to step II, 20 mg of lisinopril per day; step III, addition of 25 mg of hydrochlorothiazide (Carozide, HydroDiuril) or

5 mg of amlodipine (Norvasc) per day; or step IV, addition of 5 mg of amlodipine or 6 mg of prazosin (Minipress) per day. Patients who were not able to take angiotensin-converting enzyme inhibitors because of contraindications were given atenolol (Tenormin) (50 or 100 mg/day) instead of lisinopril. Follow-up visits were planned at 1, 2, 4, 6, 8, 10, and 12 months. Treatment decisions were made by one physician at the coordinating center who was blind to randomization of the two groups. Medical treatment was increased stepwise, left unchanged, or stepwise reduced if the diastolic blood pressure was above target (> 89 mm Hg), within the target range (80—89 mm Hg), or below target (< 80 mm Hg), respectively.

The main end points of the study were blood pressure (home, office, and ambulatory), intensity of antihypertensive drug treatment (with a treatment score summarizing all drugs with the maximal daily dose of each drug set at 1 unit), side effects (measured by a self-administered questionnaire to report the severity of 33 symptoms on a 5-point scale), adverse events, and electrocardiographic indexes for left ventricular mass (the R wave in lead aVL, the Sokolow-Lyon index,4 and the Cornell product5).

Cost-benefit analyses accounted for the physicians’ fees, the costs of antihypertensive treatment based on the rates of the Belgian health insurance system, and the costs for HBP.

SAS software, version 8.1 (SAS Institute Inc, Cary, NC) was used for all analyses. Mann-Whitney rank sum test and Student’s t test were used for between-group comparisons. Longitudinal changes in treatment modifications were compared by Kaplan-Meier survival function estimates, and proportions were compared by chi-square statistic.

Results

Of 606 patients enrolled at 60 centers, 400 (66.0%) were eligible to enter the study and were randomly assigned to either the HBP group

(n = 203) or the OBP group (n = 197). At baseline, the blood pressure values and other characteristics were no different between the two groups (table). A total of 53 patients did not finish the study, 27 (13.3%) in the HBP group and 26 (13.2%) in the OBP group. Thirty patients dropped out, 21 missed one or more follow-up visits, and two experienced an adverse event (one patient had a nonfatal stroke and one patient was diagnosed with a malignant brain tumor). Patients were followed up for a median of 350 days.

In both the OBP and HBP groups, the blood pressure decreased significantly after randomization. After 1 month of treatment, the decrease in blood pressure was no different for either treatment group, averaging 12.6/8.2 mm Hg in the OBP group and 11.5/8.1 mm Hg in the HBP group for OBP (figure 1). The decreases in blood pressure were the same for both groups after 6 months of treatment (figure 1). The decrease then became markedly greater in OBP patients than in HBP patients, with a total reduction of 22.0/14.0 mm Hg in the OBP group and 15.3/10.5 mm Hg in the HBP group. The differences in blood pressure reduction between the two treatment groups were 6.8 mm Hg systolic (P < .001) and 3.5 mm Hg diastolic (P < .001) after adjusting for body mass index, age, sex, and baseline blood pressure. Respective between-group differences were 4.9 and 2.9 mm Hg for HBP monitoring and 5.3 and 3.2 mm Hg for daytime ABP (all P < .001).

A greater proportion of HBP patients (26%) were able to permanently discontinue antihypertensive drug treatment than OBP patients (11.3%; 2.2 versus 1.0 patients per 100 followed up for 1 month; P < .001; figure 2). A greater proportion of HBP patients (38.7%) compared with OBP patients (45.1%), however, were increased to multiple drug treatments (3.3 versus 3.8 patients per 100 followed up for 1 month). The treatment score became markedly higher in the OBP group compared with the HBP group from the second follow-up visit on, with mean values of 1.47 (standard deviation [SD], 1.19) versus 1.03 (SD, 1.02) at the end of the trial (P = .001).

Using a 5-point scale, the average symptom score fell (P < .001) from 1.52 (SD, 0.36) to 1.40 (SD, 0.32) in the OBP group and from 1.60 (0.40) to 1.50 (0.41) in the HBP group over the follow-up period. Both groups had no difference in baseline-adjusted changes in the symptom score at 6 months (—0.07 versus –0.10) and at the end of the study (–0.10 versus –0.10). Thirteen patients in the OBP group and 10 patients in the HBP group had significant adverse events.

In both treatment groups, there was a significant decrease in the electrocardiographic indexes of left ventricular mass. These changes were similar in the OBP and HBP groups. The changes in electrocardiographic parameters were —0.03 mV in the OBP group versus –0.03 mV in the HBP group for the R

wave in aVL, —0.12 versus –0.09 mV for the Sokolow-Lyon index, and –13 versus –12 µV x seconds for the Cornell voltage after adjusting for sex, body mass index, age, and baseline measurements.

For 1 month of treatment, the medications cost Œ2,120 per 100 OBP patients and Œ1,688 per 100 HBP patients (P = .002). Per 100 patient-months, physicians’ fees averaged Œ1,789 for OBP patients and Œ1,510 for HBP patients (P < .001). The cost of home monitoring, however, partially equalized the possible savings related to less intensive

drug treatment and fewer physician visits in the HBP group. Compared with HBP patients, the costs were slightly greater but statistically significant for OBP patients (Œ3,875 versus Œ3,522 per 100 patient-months; P = .04).

Discussion

Our study showed that when treatment for high blood pressure was based on home monitoring, patients received less intensive drug treatment and costs were slightly lower than when treatment was based on office monitoring. There was no difference in general well-being and left ventricular mass between the HBP and OBP groups, however, and blood pressure was not as well controlled in the HBP group. Systolic and diastolic blood pressure differed by 6.8 and 3.5 mm Hg, respectively, using traditional office measurements. These differences are clinically significant for long-term prognosis, as shown by a meta-analysis of 30 clinical trials of hypertensive or high-risk patients. Over 3 to 5 years, there was a 25% to 30% increase in the risk of stroke and cardiovascular complications in the presence of a 5-mm Hg difference in systolic blood pressure.6 The results of the present study, therefore, do not support the idea that patient self-measurement at home might be a better guide to prescribing antihypertensive therapy than traditional blood pressure measurement at the physician’s office.

The widespread use of patient self-measurement is limited by the lack of prognostically validated diagnostic thresholds for the initiation or adjustment of antihypertensive therapy. Thresholds ranging from 125 to 140 mm Hg for systolic blood pressure and from 80 to 90 mm Hg for diastolic blood pressure have been suggested.7-10 Few prospective studies on the association between cardiovascular risk and self-measured blood pressure have been published. Home monitoring of blood pressure was found to be a better predictor of mortality than traditional office measurement of blood pressure in a Japanese population study.11

In our study, antihypertensive treatment was adjusted only according to diastolic blood pressure for various reasons. This criterion is used in most hypertension outcome trials.6 Diastolic blood pressure is used to assess cardiovascular risk in young and middle-aged subjects younger than 60 years of age12; it is even used to evaluate cardiovascular risk in older subjects.13 The treatment strategy would have been more complex if both systolic and diastolic measurements had been used. In both groups, treatment was

also adjusted to reach the same range of 80 to 89 mm Hg diastolic blood pressure, corresponding to the Ambulatory Blood Pressure Monitor-ing and Treatment of Hypertension (APTH) trial.1 One physician at the study center was able to suggest treatment modifications in a blind manner as a result of these design characteristics.

Conclusion

Until the benefit of incorporating home and ambulatory blood pressure monitoring into the customary care of hypertensive patients is confirmed by long-term outcome studies, office measurement by the physician is still the preferred method for diagnosis and treatment of hypertension, in agreement with current guidelines.14 When increased OPB in the physician’s office is the only detectable abnormality, or when patients with a normal OBP.