Journal of Human Hypertension (2002) 16, 865–873 & 2002 Nature Publishing Group All rights reserved 0950-9240/02 $25.00 www.nature.com/jhh ORIGINAL ARTICLE Comparison of trough effect of telmisartan vs perindopril using self blood pressure measurement: EVERESTE study

S Ragot1, A Ezzaher2, A Meunier2, M Poterre2, R Bourkaib2 and D Herpin1 1Faculte´ de Me´decine et de Pharmacie et CHU La mile´trie, Poitiers, France; 2Laboratoire GlaxoSmithKline, Marly le Roi, France

This multicentre study was aimed at comparing the less frequent with telmisartan (41%; n ¼ 85) than with trough effect of telmisartan and perindopril on diastolic perindopril (55%; n ¼ 115, P ¼ 0.005). Mean values of blood pressure (DBP), using self blood pressure SBPM were lower than office BP values, with a measurement (SBPM). A second objective was to difference of a greater importance at W0 than at W12: compare the data obtained from SBPM with those 6.6 vs 4.7 mmHg for systolic blood pressure (Po0.005) provided by automatic office BP measurement. A total and 3.2 vs 1.4 mmHg for DBP (Po0.0001). At W12, of 441 mild-to-moderate hypertensive patients were isolated office hypertension was found in 9% of the randomised to receive either telmisartan 40 mg or patients (n ¼ 37), while there were 14% of the patients perindopril 4 mg for a period of 12 weeks. Patients (n ¼ 55) with isolated home hypertension. In conclusion, whose clinic DBP remained higher than or equal to the trough effect on DBP was statistically higher with 90 mmHg at the end of the 6th week (W6) were given a telmisartan than with perindopril. SBPM values were double-dose regimen. Office BP and SBPM were per- lower than office BP values, with greater differences formed at baseline (W0), at W6 and at week 12 (W12), before than after treatment. About a quarter of the both with the same automatic device. A greater diminu- patients were found to be controlled with a method but tion of trough DBP was obtained with telmisartan not with the other one. (À6.6 7 6.7 mmHg) than with perindopril (À5.1 7 Journal of Human Hypertension (2002) 16, 865–873. 7.0 mmHg; P ¼ 0.018). Regarding clinic BP, the same doi:10.1038/sj.jhh.1001494 results were observed. Doubling dose was significantly

Keywords: self-blood pressure measurement; telmisartan; perindopril; trough effect

Introduction telmisartan and perindopril, using SBPM. The secondary objective of the study was to compare the The superiority of self blood pressure measurement data obtained from SBPM and automatic office BP (SBPM) on traditional office levels in predicting 1 measurement. This study was named EVERESTE, a target organ damage, as well as the risk of future French acronym meaning: EValuation de l’Efficacite´ cardiovascular disease and mortality, has recently RESiduelle du TElmisartan. been demonstrated.2 Other well-known advantages of this method are a better prediction of future blood pressure (BP) levels,3 a higher reproducibility,4,5 an improvement of both BP control and patient compliance6–8 and a suppression of the white coat Methods effect.9 Moreover, this method was evaluated to be Study population cost-effective.10 However, prospective trial data are based on office BP only, whereas SBPM would give Men and women aged 18 years or older with both a important and useful additional information. history of mild-to-moderate essential hypertension Accordingly, we undertook a multicentre, pro- and either inadequate BP control or treatment side spective, randomised, open study whose primary effects were eligible to enter the study. The main objective was to compare the trough effect of exclusion criteria were: a history of non response to a converting enzyme inhibitor or an angiotensin II receptor antagonist, a suspicion of a secondary Correspondence: Dr S Ragot, Faculte´ de Me´decine et de Pharmacie, 34, rue du Jardin des Plantes, BP 199, 86005 Poitiers hypertension, obstructive biliary diseases and non- Cedex, France. E-mail: [email protected] postmenopausal women without a reliable contra- Received 10 July 2002; revised and accepted 22 September 2002 ceptive method. Comparison of telmisartan and perindopril using SBPM S Ragot et al

866 At the end of a 3-week run-in placebo period, had been previously validated against the standard patients were eligible for randomisation if they had mercury sphygmomanometer according to the pro- a sitting office diastolic blood pressure (DBP) greater tocol of the British Hypertension Society.11 After than or equal to 90 mmHg and less than 110 mmHg the patient had rested for 5 min, three BP measure- and a sitting systolic blood pressure (SBP) less than ments were taken at 1-min intervals with the patient 180 mmHg. The subjects with a SBPM of poor in a sitting position. The mean of the three values quality (less than 4 days with valid measurements for DBP and SBP was considered for office BP instead of the 7 days requested) were excluded, as analysis. were those with a mean diastolic SBPM less than 85 mmHg. An additional exclusion criterion was a poor compliance to treatment. Home SBPM Home SBPM was performed over seven consecutive Study design days at three different times: at the end of the wash- out period (W0), the week before the interim visit This was a prospective, randomised, open, parallel (W6) and the week before the final visit (W12). group study that was conducted in accordance with Measurements were standardised and two series of the principles of the Declaration of Helsinki and three consecutive measurements were requested Good Clinical Practices. The protocol was approved each day: one in the morning before drug intake by the ‘Comite´ Consultatif des Personnes se preˆtant (between standing-up and breakfast) and one in the a` la Recherche Biome´dicale of Poitou-Charentes’ evening between dinner and going to bed. Record- and prior written informed consent was obtained ings were performed by the patient in the sitting from all of the subjects. After the 3-week placebo position after a 5-min rest using a printer-equipped wash-out period, during which any previous anti- semiautomatic device of the same type as that used hypertensive drug treatment was discontinued, for office BP measurement; tickets were automati- patients fulfilling inclusion criteria were randomly cally printed by the device and were pasted by the allocated to receive telmisartan 40 mg once daily or patient on an individual diary card. perindopril 4 mg once daily in the morning over a The home SBPM was validated by the investigator 12-week period. if the patient had performed a recording during four Treatment allocation was performed centrally, complete days at least. through an interactive voice response system Individual SBPM was defined for each period as (IVRS), in order to reduce the chance of introducing the mean of all available values (with the exception a bias in treatment allocation. Patients whose clinic of those of the first day). The mean of morning DBP remained greater than or equal to 90 mmHg at measurements was considered for morning SBPM the 6th week were given a double-dose regimen. and, taking place 24 h after the previous drug intake, Thus, as shown in Figure 1, four clinical visits it gave the trough BP. The mean of evening were scheduled during the study: selection visit measurements could be considered as an evaluation (W-3), randomisation visit (W0), interim visit (W6) of the peak effect of the drug and is defined as and final visit (W12). All of these visits took place in evening SBPM. Calculations of the mean value of the morning. Patients were instructed not to take the different BPs were done discarding from its study medication on the morning of a clinic visit; on determination the first day measurements as well as these days, study drug was administrated after the irrelevant values (defined as a SBP o80 or measurement of office blood pressure. 4240 mmHg and as a DBP o40 or 4140 mmHg). The main efficacy criterion was trough DBP (morning measurement) obtained after a 12-week Office BP measurements treatment period adjusted on baseline DBP. The secondary criteria were evening DBP, morning BP was measured at each clinic visit using a and evening SBP, normalisation rates at office semiautomatic device (OMRON 705 CP), which (DBP o90 mmHg and SBP o140 mmHg) and with SBPM (DBPo85 mmHg and SBPo135 mmHg12), and correlation coefficients between office and self-measurements.

Statistical analysis It was calculated that a total of 420 randomised patients was necessary to provide 80% power to detect a 3 mmHg (s.d. ¼ 10) difference in the final trough DBP evaluated by SBPM assuming a 20% dropout rate and two-sided testing at an overall Figure 1 Protocol design. alpha significance level of 0.05.

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

867 Statistical analysis was conducted with SAS (SAS the perindopril group). Figure 2 summarises the Institute, Cary, NC, USA). reasons of exclusion from the per-protocol popula- Analysis of variance covariance for repeated tion. Considering the ITT population, patients’ measures was performed with Proc mixed to mean age was 55 7 12 years (range 27–89 years), compare the W6 and W12 BP values or changes 55% were men and 97% were Caucasians. The from baseline between the two treatment groups. median duration of hypertension was 1 year, and the Data are given as mean values 7 s.d. The values of mean BPs measured with automatic device at office the between-group differences (telmisartan minus at baseline was 158 7 13 and 98 7 6 mmHg for SBP perindopril) in BP reduction from baseline to W12 and DBP, respectively. The two treatment groups are given adjusted for baseline BP level and with were comparable with respect to all baseline their 95% confidence interval (CI). The differences characteristics (Table 1). in the rates of normalisation under treatment were assessed using the w2 test. A comparison of office BP and SBPM was done Efficacy according to the method of Bland and Altman:13 the average of the measurements evaluated by both As shown in Figure 3, trough DBP was significantly methods was plotted against their difference, for lower in the telmisartan group than in the perindo- both SBP and DBP. pril group when evaluated either with SBPM or Spearman’s correlation coefficients were also with clinic measurement and both at W12 and W6. calculated. We report only analyses done by inten- The same findings were obtained with trough SBP. tion to treat (ITT). Statistical significance was set at Moreover, the trough pulse pressure (PP) was Po0.05 (two-sided). significantly lower in the telmisartan group than in the perindopril group at W12, when evaluated by SBPM only. Regarding evening SBPM, there was no Results significant difference between telmisartan and perindopril (data not shown). A total of 671 patients attending 105 outpatient The average decrease in trough DBP (value of French centres were selected. After the 3-week week 12 minus value of baseline) was À6.6 mmHg placebo phase before randomisation, 441 met the for telmisartan and À5.1 mmHg for perindopril inclusion/exclusion criteria and were randomised to when evaluated with SBPM. Accordingly, the receive either telmisartan or perindopril. The ITT adjusted between-treatment difference was population included 435 of the original 441 patients À1.4 mmHg (95% CI ¼ [À2.74; À0.14 mmHg]; (five patients, two in the telmisartan group and three P ¼ 0.03). in the perindopril group were excluded because of The clinic method brought to similar results with no-assessment of home BP and office BP on active a diminution in diastolic from baseline of drug; furthermore, one patient did not receive the À8.8 mmHg with telmisartan and À6.3 mmHg with study treatment). perindopril, that is, mean difference between the two In all, 368 patients took part in per-protocol treatment groups of À2.5 mmHg (95% CI ¼ [À4.11; analysis (188 in the telmisartan group and 180 in À0.89]; P ¼ 0.002).

Figure 2 Trial profile.

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

868 Table 1 Patient demographics and baseline characteristics (ITT population)

Telmisartan n=217 Perindopril n=218

Men/women (no; %) 124 (57%)/93 (43%) 114 (52%)/104 (48%) Age (years) 55.1 7 11.6 55.5 7 12.0 Caucasian (no; %) 207 (97%) 213 (98%) Obesity (no; %) 52 (24%) 59 (27%) Tobacco Current smoker (no; %) 40 (18%) 42 (19%) Former smoker (no; %) 52 (24%) 52 (24%) Nonsmoker (no; %) 125 (58%) 124 (57%) History of CV events (no; %) 28 (13%) 30 (14%) Type II diabetes (no, %) 12 (6%) 15 (7%) Duration of hypertension o1 year (no, %) 74 (34%) 70 (32%) 1–10 years (no, %) 106 (49%) 103 (47%) X10 years (no, %) 36 (17%) 45 (21%) Family history of hypertension (no, %) 134 (62%) 127 (58%) Antihypertensive treatment (no, %) 116 (53%) 120 (55%) Monotherapy (no, %) 90 (78%) 92 (77%) X2 drugs (no, %) 26 (22%) 28 (23%) Baseline OBPM (mmHg) 158 7 13/98 7 6 159 7 13/98 7 6 Baseline SBPM (mmHg) 152 7 15/94 7 6 153 7 15/94 7 6

CV: cardiovascular; OBPM: office blood pressure measurement; SBPM: self blood pressure measurement. All comparisons were nonsignificant.

Figure 3 Blood pressure changes (mean 7 s.e.m.) induced by telmisartan (–~–) and perindopril ( )–ITT Population. Left panel: SBPM; right panel: OBPM; W0: randomisation visit; W6: interim visit; W12: final visit. Comparison of telmisartan vs perindopril:*Po0.05; **Po0.01; ***Po0.005.

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

869 Table 2 Comparison of the normalisation rates with telmisartan Agreement between the two methods and perindopril according to both methods of measurement (intent-to-treat population) Table 3 provides information about the agreement between the two methods of measurement. Con- Telmisartan Perindopril sidering the overall ITT population, mean values of SBP, DBP and PP obtained by SBPM were signifi- Trough DBP cantly lower than those obtained by automatic office OBPM o90 mmHg: n(%) W6 127 (59%)*** 95 (44%) BP both at W0 and W12. However, there was a weak W12 122 (58%)** 96 (46%) but nonetheless statistically significant linear corre- SBPM o85 mmHg: n(%) lation between the two methods of measurement. W6 96 (45%)** 68 (32%) The stronger correlation was observed for PP at W12 95 (47%)* 70 (36%) W12. For SBP and DBP, the mean differences Trough SBP between the methods were significantly lower at OBPM o140 mmHg: n(%) W12 than at W0 (Ppaired testo0.005 for SBP and W6 82 (38%)** 56 (26%) o0.0001 for DBP). The 95% CI of the between- W12 97 (46%)*** 67 (32%) method differences was large: always greater SBPM o135 mmHg: n(%) W6 79 (37%) 69 (33%) than 30 mmHg and of more than 50 mmHg for SBP W12 85 (42%) 79 (40%) at W0. Figure 4 represents the comparison of both DBP: diastolic blood pressure; SBP: systolic blood pressure; OBPM: measurement methods for DBP at W0 and at W12 office blood pressure measurement; SBPM: self blood pressure according to the method of Bland and Altman. measurement; W6: interim visit; W12: final visit. Comparisons of telmisartan vs perindopril: *Po0.05, **Po0.01, Figure 5 shows the distribution of normalised and ***Po0.005. non-normalised patients according to both methods. The discrepancies between the two methods were balanced for SBP and this is not the case for DBP where the most frequent DBP discrepancies were Similar findings were obtained with trough SBP: related to patients who were normalised according adjusted mean difference evaluated with SBPM to office measurement and not normalised according was À3.4 mmHg in favour of telmisartan (95% to SBPM. CI ¼ [À5.69; À1.08]; P ¼ 0.004). As to clinic SBP, At the end of the study, 37 patients were found the adjusted mean difference was À3.4 mmHg (95% with isolated office hypertension, that is, 9% of our CI [À6.24; À0.64]; P ¼ 0.016). population study, while there were 55 patients, that Regarding trough PP, the between-treatment dif- is, 14%, with isolated home hypertension. ference obtained from SBPM was significant: –2 mmHg (95% CI ¼ [À3.44; À0.54]; P ¼ 0.007). By contrast no significant difference was found with Feasibility clinic measurement: À1 mmHg (95% CI ¼ [À3.17; 1.14]; NS). The rate of patients excluded from ITT population Normalisation rates are given in Table 2. The DBP for insufficient SBPM quality was 3.7% (nine normalisation rates with telmisartan were higher patients in the telmisartan group and seven patients than those with perindopril at W12 and at W6 in the perindopril group) before W0 and 9.9% (18 whatever the method of measurement. As a con- patients in the telmisartan group and 25 in the sequence, doubling dose was significantly less perindopril group) before W12. A poor quality of frequent with telmisartan (41%; n ¼ 85) than with SBPM resulted in 81% of the exclusions from per- perindopril (55%; n ¼ 115, P ¼ 0.005). protocol population.

Table 3 Agreement between the two methods of BP measurement (ITT population)

Trough BP SBP (W0) DBP (W0) PP (W0) SBP (W12) DBP (W12) PP (W12)

Office BP/SBPM: 158.4 7 13.0/ 97.7 7 5.8/ 60.7 7 12.3/ 146.0 7 17.0/ 90.0 7 9.0/ 55.9 7 13.3/ (Mean 7 s.d.) (mmHg) 151.9 7 16.3** 94.5 7 7.1* 57.4 7 13.4** 141.3 7 18.2** 88.6 7 8.5* 52.7 7 13.3**

Correlation coefficient 0.63** 0.43** 0.63** 0.70** 0.54** 0.78** Mean difference 6.6 3.2 3.4 4.7 1.4 3.3 (OBPMÀSBPM) (mmHg) Limits of agreement [À19.0; 32.3] [À10.3; 16.8] [À18.3; 25.1] [À20.4; 29.8] [À14.1; 16.9] [À13.8; 20.5] (Mean 7 1.96s.d.) (mmHg)

BP: blood pressure; SBP: systolic blood pressure; DBP: diastolic blood pressure; PP: pulse pressure; OBPM: office blood pressure measurement; SBPM: self blood pressure measurement; s.d.: standard deviation; W0: randomization visit; W6: interim visit; W12: final visit. Comparisons of telmisartan vs perindopril: *Po0.0005; **Po0.0001.

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

870

Figure 4 Comparison of diastolic SBPM and diastolic OBPM according to the method of Bland and Altman. Upper panel: at randomisation visit; lower panel: at final visit. OBPM: office blood pressure measurement; SBPM: self blood pressure measurement.

Table 4 gives, for each visit and for all of the visits, Adverse events the number of patients who performed a complete (six measurements per day) SBPM over a total of The overall incidence of adverse events was com- 7 days or 4 days. Considering independently each parable in both treatment groups: 34% (n ¼ 74) visit, about one-half of the patients was able to with telmisartan and 32% (n ¼ 70) with perindopril. record BP for 7 days and less than 30% were able to Most of them were of mild-to-moderate intensity do so for all visits. The sequence of 4 days showed a and transient. As expected, the occurrence of greater feasibility: accordingly, almost 90% of the cough was more frequent with perindopril: 5% patients recorded a complete sequence of 4 days at (n ¼ 12) than with telmisartan: o1% (n ¼ 2), each of the three visits. P ¼ 0.007.

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

871 greater efficacy of telmisartan, in comparison with perindopril. The antihypertensive efficacy of telmisartan has been shown in clinical studies using measurement office BP, to be at least equivalent to that of angiotensin-converting enzyme inhibitors, namely: lisinopril in mild-to-moderate hypertensive pa- tients,14 enalapril in mild-to-moderate hyper- tensives,15,16 in severe hypertensives17 and in elderly patients.18 Interestingly, telmisartan was described by Mal- lion et al19 to be more effective in lowering SBP and DBP during the 18 to 24-h postdose period than was losartan using ambulatory BP monitoring.19 Like- wise, a greater reduction in ambulatory DBP was found with telmisartan than with amlodipine during the night time interval and the last 4 h of the design period in a study by Lacourcie`re et al.20 Figure 5 Distribution of normalised patients and nonnormalised Of note, doubling the dosage of both telmisartan patients at the final visit (W12) according to both BP measurement and perindopril resulted in a very low increase in methods (n ¼ 397 patients). &, patients nonnormalised according the number of normalised patients. Obviously, in to both methods: &, patients normalised according to both clinical practice, the most efficient strategy in the methods; , patients normalised according to office measure- ment only (isolated home hypertension); , patients normalised presence of a poor response to monotherapy should according to SBPM only (isolated office hypertension). be not to increase the dose, but to add a diuretic or to shift to another drug, acting through a different mechanism. Accordingly, Lacourcie`re et al21 Table 4 Feasability of SBPM: number of patients with a sequence showed that a ‘telmisartan 80 mg/HCTZ 12.5 mg (six measurements per day) of 7 days and of 4 days at W0, W6 and fixed dose combination confers significant addi- W12 (ITT population) tional BP reductions compared with continuation of telmisartan monotherapy in nonresponders’ (41.5% Telmisartan Perindopril of patients with normalised BP in the combination Patients with 7 days of complete SBPM treatment group vs 26.1% in telmisartan group; W0 111 (52%) 128 (60%) Po0.05), Furthermore, in a recent crossover study W6 108 (51%) 98 (46%) by Dickerson et al,22 a significant correlation was W12 104 (51%) 90 (46%) found between the BP responses to angiotensin- W0+W6+W12 52 (26%) 50 (26%) converting enzyme inhibitors and betablockers, as Patients with 4 days of complete SBPM well as between those to calcium channel blockers W0 202 (94%) 206 (96%) and diuretics, but not between the other four pairing W6 198 (93%) 194 (92%) of treatments. Then it could be speculated that W12 192 (94%) 187 (95%) patients who fail to respond to renin–angiotensin W0+W6+W12 175 (87%) 169 (87%) system blockade should be subsequently given a calcium channel blocker or a diuretic (and con- SBPM: self blood pressure measurement; W0: randomization visit; W6: interim visit; W12: final visit. versely).

Comparison of office and home BP measurements Discussion Considering the average BP values, we found along Antihypertensive effect of telmisartan and of with other authors that home BP levels were lower perindopril than their corresponding office values with a difference of 1.4 mmHg for DBP and of 4.7 mmHg In the present study, telmisartan 40 mg produced for SBP at the end of the study. Before treatment clinically relevant decreases in trough DBP, SBP and (visit W0) these differences were larger: 3.2 mmHg PP that were slightly but significantly greater than for DBP and 6.6 mmHg for SBP. These findings are in those produced by perindopril both with SBPM and keeping with the existing published data: with automatic office measurement. Categorical analysis of DBP and SBP rates of normalisation * In a previous study23 aimed at comparing the confirmed these differences in favour of telmisartan. antihypertensive effects of trandolapril and losar- However, considering the absence of statistical tan in 229 hypertensives, we found a difference of difference in evening BP between the two treatment 1.3 and 8.9 mmHg for DBP and SBP, respectively, groups, these data have to be understood as a more under treatment and of 11.2 and 17.5 mmHg sustained antihypertensive effect rather than as a before treatment. The between-method differences

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

872 were of a greater magnitude in this previous study prognosis of patients with normal clinic BP and than they are in the present study, a possible elevated ambulatory or self-measured BP. Of note, explanation could be that office BP was measured we found in our study an unexpected high rate of using a mercury sphygmomanometer, not an patients with isolated home BP at the end of the automatic device. study. A cross-sectional study28 has suggested these * Likewise, pooling the results of 12 studies using patients to be at a higher cardiovascular risk, as long home BP monitoring, Brook24 reported mean as they have been found with more risk factors and a home BP values of 4.6 mmHg (95% CI ¼ [1.8; more frequent history of cardiovascular diseases, 7.4]) and 6.5 mmHg (95% CI ¼ [1.3; 13.2]) lower in comparison with other patients. However, we than office values for DBP and SBP, respectively. should have to wait for 3 years to know whether this * In contrast, within a subset of the HOT study hypothesis is confirmed or not, by a longitudinal population including more than 900 patients, approach. Kjeldsen et al25 has reported very slight differ- ences between office and home values of about 0.2 mmHg for DBP and 0.5 mmHg for SBP, at a Feasibility time when the titration of antihypertensive medication had been finalised. Despite the small SBPM feasibility is currently considered as good, size of the mean differences, the 95% limits of even if it has been said that SBPM technique needs a agreement between the two methods were very greater physician involvement and a selection of large, as compared with those of the present appropriate patients to comply with the SBPM study: À17.4 to 17.9 mmHg for DBP and À29.5 to protocol. Our study confirms the good feasibility 30.5 mmHg for SBP. In addition, the correlation of this technique; however, a SBPM poor quality between the two methods was rather low: 0.35 for recording was the reason why 16 and 43 patients DBP and 0.45 for SBP, vs 0.54 and 0.70 in our (81%) had to be excluded from the per-protocol study. The differences between the two studies population, before randomisation and before the could be partly because of the conditions of end of the study, respectively. measurement both at home and in the office: (1) The monitoring schedule defined for this study the semiautomatic device used in the HOT study consisted of six measurements per day over a total of was not print-equipped; (2) measurements were 7 days: such a design is not in agreement with the not standardised for intake of antihypertensive most recent recommendations, which suggest to medication; (3) the morning measure had to be perform four to six measurements a day over a 3 to 4- 12 done before leaving home and the evening day period. determination was required in the afternoon after returning home. Obviously, the findings of Kjeld- sen’s substudy, as well as those of the present Study limitations study, point out the irrelevance of an individual The main limitation of this study is related to the extrapolation of office BP measurement to home lack of blinded drug allocation. Such an open-label BP readings. design has been arbitrarily chosen. However, BP measurements were performed under standardised Concerning the attenuation of the difference conditions and using automatic machines, which between office BP and SBPM with drug intake, it should avoid bias in the primary end point of the could be interesting to refer to a study by Parati 26 study. In contrast, the knowledge of drug allocation et al, related to the effects of calcium channel could have resulted in an impact of tolerability blockers and ACE inhibitors vs placebo, on the assessment. Interestingly, in a double-blind study, a differences between clinic BP measurement and cough incidence of 6.5% has been reported in ambulatory BP monitoring. In their experience, the patients receiving telmisarten vs 15.8% in those greater the baseline between-method difference, the receiving enalapril;29 in another randomised double- greater the treatment-induced reduction in this blind study, a treatment-related cough was found in difference; furthermore, they showed placebo to none of the patients who were given 40–80 mg of have an effect of the same magnitude than did telmisartan, whereas it occurred in 4.2% of patients pharmacological drugs on the clinic–ambulatory on enalapril and in 1.3% of patients on placebo.16 difference, suggesting either a habituation of the patients to clinic BP measure, or a regression towards the mean, or both. Of course, the same explanations could be given as to the differences References between clinic BP and SBPM. Accordingly, Imai 27 1 Abe H et al. Relation of office and home blood pressure et al considered the magnitude of regression to left ventricular hypertrophy and performance in towards the mean to be rather small for SBPM. patients with hypertension. High Blood Press 1992; 1: While a large body of literature has demonstrated 279–285. the lack of prognostic significance of the white coat 2 Ohkubo T et al. Home blood pressure measurement has effect, there are still no clear information about the a stronger predictive power for mortality than does

Journal of Human Hypertension Comparison of telmisartan and perindopril using SBPM S Ragot et al

873 screening blood pressure measurement: a population- 17 Neutel JM, Smith DH, Reilly PA. The efficacy and based observation in Ohasama, Japan. J Hypertens safety of telmisartan compared to enalapril in patients 1998; 16: 971–975. with severe hypertension. Int J Clin Pract 1999; 53: 3 Nesbitt SD et al. Home blood pressure as a predictor 175–178. of future blood pressure stability in borderline hyper- 18 Karlberg BE, Lins LE, Hermansson K. Efficacy and tension. The Tecumseh study. Am J Hypertens 1997; safety of telmisartan, a selective AT1 receptor antago- 10: 1270–1280. nist, compared with enalapril in elderly patients with 4 James JD, Pickering TG, Yees LS. Reproducibility primary hypertension. TEES Study Group. J Hypertens of average ambulatory, home and clinic pressures. 1999; 17: 293–302. Hypertension 1998; 11: 545–549. 19 Mallion JM, Siche´ JP, Lacoucie`re Y and the Telmisartan 5 Sakuma M et al. Reproducibility of home blood Blood Pressure Monitoring Group. ABPM comparison pressure measurements over a one-year period. Am J of the antihypertensive profiles of the selective angio- Hypertens 1997; 10: 798–803. tensin II receptor antagonists telmisartan and losartan 6 Carnahan JE, Nugent CA. The effects of self-monitoring in patients with mild-to-moderate hypertension. J by patients on the control of hypertension. Am J Med Hum Hypertens 1999; 13: 657–664. Sci 1975; 269: 69–73. 20 Lacourcie`re Y et al. A comparison of the efficacy and 7 Zarnke K, Feagan B, Mahon J, Feldman RD. A duration of action of the angiotensin II receptor blocker randomized study comparing a patient-directed hyper- telmisartan to amlodipine. Blood Press Monit 1998; 3: tension management strategy with usual office-based 295–302. care. Am J Hypertens 1997; 10: 58–67. 21 Lacourcie`re Y et al. Efficacy and tolerability of a fixed- 8 Edmonds D et al. Does self-measurement of blood dose combination of telmisartan plus hydrochlorothia- pressure improve patient compliance in hypertension? zide in patients uncontrolled with telmisartan mono- J Hypertens 1985; 3 (Suppl): S31–S34. therapy. J Hum Hypertens 2001; 15: 763–770. 9 Pickering T. Recommendations for the use of home 22 Dickerson JE et al. Optimisation of antihypertensive (self) and ambulatory blood pressure monitoring. treatment by crossover rotation of four major classes. American Society of Hypertension Ad Hoc Panel. Am Lancet 1999; 353: 2008–2013 J Hypertens 1996; 9: 1–11. 23 Ragot S, Genes N, Vaur L, Herpin D. Comparison of 10 Soghikan K et al. Home blood pressure monitoring: three blood pressure measurement methods for the effect on use of medical services and medical care evaluation of two antihypertensive drugs: feasibility, costs. Med Care 1992; 30: 855–865. agreement, and reproducibility of blood pressure 11 O’Brien E, Mee F, Atkins N, Thomas M. Evaluation of response. Am J Hypertens 2000; 13: 632–639. three devices for self-measurement of blood pressure 24 Brook RD. Home blood pressure: accuracy is indepen- according to the revised British Hypertension dent of monitoring schedules. An J Hypertens 2000; 13: Society Protocol: the Omron HEM-705CP, Philips 625–631. HP5332, and Nissei DS-175. Blood Press Monit 1996; 25 Kjeldsen SE et al. Hypertension optimal treatment 1: 55–61. (HOT) study. Home blood pressure in treated hyper- 12 Asmar R, Zanchetti A. Guidelines for the use of self- tensive subjects. Hypertension 1998; 31: 1014–1020. blood pressure monitoring: a summary report of the 26 Parati G, Omboni S, Mancia G. Difference between First International Consensus Conference. Groupe office and ambulatory blood pressure and response Evaluation & Measure of the French Society of to antihypertensive treatment. J Hypertens 1996; 14: Hypertension. J Hypertens 2000; 18: 493–508. 791–797. 13 Bland JM, Altman DG. Statistical methods for asses- 27 Ima Y et al. Usefulness of home blood pressure sing agreement between two methods of measurement. measurements in assessing the effect of treatment in Lancet 1986; I: 307–310. a single-blind placebo-controlled open trial. J Hyper- 14 Neutel JM et al. Comparison of telmisartan with tens 2001; 19: 179–185. lisinopril in patients with mild-to-moderate hyperten- 28 Bobrie G et al. Is isolated home hypertension as sion. Am J Ther 1999; 6: 161–166. opposed to isolated office hypertension a sign of 15 Smith DH, Matzek KM, Kempthorne-Rawson J. Dose greater cardiovascular risk? Arch Intern Med 2001; response and safety of telmisartan in patients with 161: 2205–2211. mild to moderate hypertension. J Clin Pharmacol 2000; 29 Karlberg BE, Lins LE, Hermansson K, for the TEES 40: 1380–1390. Study Group. Efficacy and safety of telmisartan, a 16 Smith DH, Neutel JM, Morgenstern P. Once daily selective AT1 receptor antagonist, compared with telmisartan compared with enalapril in the treatment enalapril in elderly patients with primary hyperten- of hypertension. Adv Ther 1998; 15: 229–240. sion. J Hypertens 1999; 17: 293–302.

Journal of Human Hypertension