On-Treatment Diastolic Blood Pressure and Prognosis in Systolic Hypertension

ORIGINAL INVESTIGATION On-Treatment Diastolic Blood Pressure and Prognosis in Systolic Hypertension

Robert H. Fagard, MD; Jan A. Staessen, MD; Lutgarde Thijs, MSc; Hilde Celis, MD; Christopher J. Bulpitt, MD; Peter W. de Leeuw, MD; Gastone Leonetti, MD; Jaakko Tuomilehto, MD; Yair Yodfat, MD

Background: It has been suggested that low diastolic mortality, but not cardiovascular mortality, increased with blood pressure (BP) while receiving antihypertensive treat- low diastolic BP with active treatment (PϽ.005) and with ment (hereinafter called on-treatment BP) is harmful in placebo (PϽ.05); for example, hazard ratios for lower older patients with systolic hypertension. We examined diastolic BP, that is, 65 to 60 mm Hg, were, respectively, the association between on-treatment diastolic BP, mor- 1.15 (95% confidence interval, 1.00-1.31) and 1.28 (95% tality, and cardiovascular events in the prospective pla- confidence interval, 1.03-1.59). Low diastolic BP with ac- cebo-controlled Systolic Hypertension in Europe Trial. tive treatment was associated with increased risk of cardiovascular events, but only in patients with coronary Methods: Elderly patients with systolic hypertension heart disease at baseline (PϽ.02; hazard ratio for BP 65-60 were randomized into the double-blind first phase of the mm Hg, 1.17; 95% confidence interval, 0.98-1.38). trial, after which all patients received active study drugs (phase 2). We assessed the relationship between out- Conclusions: These findings support the hypothesis that come and on-treatment diastolic BP by use of multivar- antihypertensive treatment can be intensified to pre- iate Cox regression analysis during receipt of placebo vent cardiovascular events when systolic BP is not un- (phase 1) and during active treatment (phases 1 and 2). der control in older patients with systolic hypertension, at least until diastolic BP reaches 55 mm Hg. However, a Results: Rates of noncardiovascular mortality, cardio- prudent approach is warranted in patients with concomi- vascular mortality, and cardiovascular events were 11.1, tant coronary heart disease, in whom diastolic BP should 12.0, and 29.4, respectively, per 1000 patient-years with probably not be lowered to less than 70 mm Hg. active treatment (n=2358) and 11.9, 12.6, and 39.0, respectively, with placebo (n=2225). Noncardiovascular Arch Intern Med. 2007;167(17):1884-1891

HE EXISTENCE OF A J-CURVE tension and isolated systolic hypertension. relationship between mor- Therefore, the results may not apply to pa- tality or cardiovascular event tients with systolic hypertension who may and diastolic blood pres- be particularly vulnerable to low on- Author Affiliations: sure (BP) during antihyper- treatment diastolic BP, as shown in a post Hypertension and tensive treatment (hereinafter called on- hoc analysis of the Systolic Hypertension Cardiovascular Rehabilitation T Unit, University of Leuven, treatment BP) remains a matter of debate in the Elderly Program (SHEP), in which 1 Leuven, Belgium (Drs Fagard, in patients with hypertension. Apart from a J-curve was observed for incident car- Staessen, and Celis and intervention trials, in which patients are diovascular events during active treat- Ms Thijs); Imperial College, randomized to different target BP levels,2 ment but not during administration of pla- Hammersmith Hospital, placebo- or nontreatment-controlled trials cebo.15 The SHEP investigators concluded London, United Kingdom are probably the best way to evaluate the that it could be harmful to intensify anti- (Dr Bulpitt); University of J-curve. A large meta-analysis of indi- hypertensive therapy when diastolic BP has Maastricht, Maastricht, the 3-10 Netherlands (Dr de Leeuw); vidual patient data from 7 trials showed reached 70 mm Hg. In addition, the SHEP Istituto Auxologico Italiano, a higher incidence of cardiovascular death analysis did not stratify for coronary heart Ospedale San Luca, Milan, Italy with lower on-treatment diastolic BP, but disease (CHD) at baseline, whereas the risk (Dr Leonetti); National Public this relationship was also observed for non- of low on-treatment diastolic BP could be Health Institute and the cardiovascular death and similar relation- particularly high in patients with CHD.1,16 University of Helsinki, Helsinki, ships existed in the control patients; thus, Because of the far-reaching conse- Finland, and South the worse prognosis of lower diastolic BP quences of the conclusions from the SHEP Ostrobothnia Central Hospital, could not be solely ascribed to antihyper- trial, we analyzed the relationship be- Seina¨joki, Finland tensive treatment. A more recent meta- tween on-treatment diastolic BP, mortal- (Dr Tuomilehto); and 11 4-10,12-14 Department of Family analysis involving 10 trials con- ity, and cardiovascular events in older Medicine, Hadassah Medical cluded that lowering diastolic BP to less patients with systolic hypertension ran- School, Hebrew University of than 70 mm Hg did not cause harm. How- domized into the Systolic Hypertension in Jerusalem, Jerusalem, Israel ever, these meta-analyses did not differ- Europe (Syst-Eur) Trial12,17 and also ac- (Dr Yodfat). entiate between systolic-diastolic hyper- counted for CHD at baseline.

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1884

Downloaded From: https://jamanetwork.com/ on 09/30/2021 METHODS Table 1. Patient Characteristics at Baseline in the Placebo and Active Treatment Groupsa TRIAL DESIGN Active 12 The protocol of the Syst-Eur Trial was approved by the ethics Placebo Treatment committees of the University of Leuven, Leuven, Belgium, and Group Group participating centers; all subjects in the 198 centers gave in- Characteristic (n = 2225) (n = 2358) formed consent. Eligible patients had to be 60 years or older. Dur- Age, y 70.1 ± 6.6 70.2 ± 6.7 ing the run-in period with placebo, patients were seen at 3 base- Female sex 66.5 67.4 line visits 1 month apart. Data about medical history were collected, Body mass indexb 27.0 ± 4.0 27.0 ± 4.2 including history of CHD, which comprised myocardial infarc- Blood pressure, mm Hg tion and angina pectoris, based on supporting documents and ex- Systolic 173.9 ± 10.1 173.7 ± 9.8 aminations available to the local investigators. At each visit, BP Diastolic 85.5 ± 5.9 85.5 ± 5.8 was measured twice with a standard sphygmomanometer with Pulse rate, beats/min 73.0 ± 8.1 73.2 ± 7.9 the patient sitting. Patients were admitted to the double-blind phase Previous antihypertensive therapy 46.7 46.3 of the trial when they had an average run-in period systolic BP Previous cardiovascular complicationsc 30.3 29.4 of 160 to 219 mm Hg with diastolic BP less than 95 mm Hg. History of coronary heart diseasec 14.7 14.2 Patients were randomized to active medication or matched Diabetes 10.1 10.5 placebo groups. Active treatment consisted of nitrendipine (10- Current smoking 7.0 7.3 40 mg/d), which could be combined with or replaced by enala- Serum creatinine, mg/dL 1.0 ± 0.2 1.0 ± 0.2 pril maleate (5-20 mg/d) or hydrochlorothiazide (12.5-25 mg/d) Serum cholesterol, mg/dL 236 ± 46 231 ± 46 or both drugs, to reduce the sitting systolic BP by 20 mm Hg or more and to less than 150 mm Hg. At each 3-month visit, BP was SI conversion factors: To convert creatinine to millimoles per liter, multiply measured twice with the patient sitting and the 2 BP measure- by 88.4; cholesterol to millimoles per liter, multiply by 0.0259. a Data are given as mean ± SD or percentage. ments were averaged. When phase 1 of the Syst-Eur Trial was b Calculated as weight in kilograms divided by height in meters squared. stopped on February 14, 1997, after a median follow-up of 2.0 c Includes coronary heart disease (myocardial infarction and angina pectoris), years (range, 1 month to 8.1 years),12 patients in the control group heart failure, left ventricular hypertrophy, stroke, central nervous system. were switched to the active study treatment regimen and patients involvement, renal involvement, retinopathy grade 3, and aortic aneurysm. in the active treatment group continued receiving active treatment (phase 2).17 Visits were scheduled every 6 months in the study centers and the same information was collected as during the double- ables with potential influence on BP and outcome, that is, age, blind phase of the trial. The extended follow-up ended on Decem- sex, smoking status, and previous antihypertensive treatment at ber 31, 2001. In the present analyses, follow-up data in the baseline, and weight, stroke, myocardial infarction, and diabetes placebo group are from phase 1 of the trial and follow-up data in mellitus both at baseline and as time-dependent covariates dur- the active treatment group are from phases 1 and 2. For patients ing follow-up. The analyses were performed with and without sys- who withdrew from the study or who could not be followed up tolic BP as a continuous time-dependent covariate. We plotted as planned, investigators collected data about vital status, occur- the adjusted hazard ratio (HR) vs on-treatment diastolic BP. The rence of diseases, and use of antihypertensive medication via yearly (adjusted HR−1)ϫ100 is the percent increase in risk associated telephone contact with patients, family members, or general 12,17 with the 5–mm Hg lower on-treatment diastolic BP. In the sec- practitioners. ond analysis, we calculated HRs in 5–mm Hg decrements of on- treatment diastolic BP, as in the SHEP study.15 For each cutoff OUTCOMES value, outcome was compared between patients with diastolic BP below that level and patients with diastolic BP at or above that Outcome measures were as follows: (1) fatal and nonfatal cere- level. Adjustment was performed as described, with inclusion of brovascular events including stroke and transient ischemic attack; on-treatment systolic BP as a dichotomized variable (cutoff value, (2) fatal and nonfatal CHD events including sudden death, myo- 140 mm Hg), as in the SHEP trial.15 These analyses were re- cardial infarction, and coronary revascularization (percutane- peated with on-treatment systolic BP as a continuous variable and ous coronary interventions and coronary artery bypass surgery); without systolic BP. All analyses were performed on the basis of (3) an aggregate of all cardiovascular events (cardiovascular death, intention to treat. A 2-tailed PՅ.05 was considered statistically myocardial infarction, stroke, heart failure, coronary revascular- significant. ization, aortic aneurysm, and transient ischemic attack); (4) cardiovascular mortality including all fatal cardiovascular events; and (5) noncardiovascular death. Events were corroborated by the Syst- RESULTS Eur Trial End Point Committee.12,17 PATIENT CHARACTERISTICS STATISTICAL ANALYSIS AT BASELINE

Database management and statistical analyses were performed Among the 4695 patients randomized in phase 1 of the using SAS software (version 8.2; SAS Institute Inc, Cary, North Syst-Eur Trial, 2225 patients in the placebo group and Carolina). Data are reported as mean±SD or as percentage. The 2358 patients in the active treatment group had at least prognostic value of diastolic BP during follow-up was analyzed 1 follow-up visit. Patient characteristics at baseline were in patients with at least 1 follow-up visit. In the first analysis, diastolic BP (linear and quadratic term) was entered as a time- not significantly different between the groups (Table 1). dependent covariate in multivariate Cox regression models.18 In Mean±SD age of the total study population was 70.2±6.7 this analysis, the likelihood of an event that occurred at time years; 66.9% were women. Mean±SD BP was 173.8±9.9/ t depends on the value of the last available diastolic BP before time 85.5±5.8 mm Hg; 46.5% of patients had been treated with t for all subjects still in follow-up at time t. We adjusted for vari- antihypertensive drugs in the last 6 months before in-

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1885

Downloaded From: https://jamanetwork.com/ on 09/30/2021 the active treatment group and 241 patients in the pla- Table 2. Frequency Distribution of Mean Diastolic cebo group received open-label antihypertensive drugs only. Blood Pressure and All Diastolic Blood Pressure The remaining 131 patients in the active treatment group Measurements During Follow-up in the Placebo and Active Treatment Groupsa and 202 patients in the placebo group were untreated. Mean systolic BP, that is, the mean systolic BP of all postbaseline Active Treatment results for each patient, was 149.0±10.3 mm Hg (range,116- Placebo Group Group 206 mm Hg) during active treatment and 161.8±13.6 mm Hg (range, 114-238 mm Hg) during placebo admin- Mean All Mean All istration. These values were, respectively, 78.8±6.0 mm Hg Variable DBP DBPs DBP DBPs (range, 50-100 mm Hg) and 83.6±6.4 mm Hg (range, 50- No. of BP 2225 21 122 2358 41 774 110 mm Hg) for diastolic BP. Table 2 gives the distribu- measurements DBP category, mm Hg tion of mean diastolic BP and of all diastolic BP measure- Ն95 3.4 7.0 0.4 2.1 ments in the 2 groups. The percentage of patients with at Ͻ95-85 41.6 40.4 13.9 20.1 least 1 diastolic BP measurement less than 55 mm Hg dur- Ͻ85-75 47.2 38.6 63.7 48.0 ing follow-up was 5.7% in the active treatment group and Ͻ75-65 7.3 12.0 20.3 24.0 1.1% in the placebo group. The rates of events per 1000 Ͻ 65-55 0.4 1.7 1.5 5.0 patient-years for cardiovascular events, cardiovascular mor- Ͻ55 0.04 0.2 0.2 0.7 tality, and noncardiovascular mortality were 39.0, 12.6, and Abbreviations: BP, blood pressure; DBP, diastolic blood pressure. 11.9, respectively, during 5219 patient-years of follow-up a Blood pressure values are given as percentage. during placebo treatment and 29.4, 12.0, and 11.1 during 14 511 patient-years during active treatment (Table 3).

Table 3. Numbers and Rates per 1000 Patient-Years PROGNOSTIC SIGNIFICANCE of First-Occurring Events in the Placebo OF DIASTOLIC BP: CONTINUOUS ANALYSIS and Active Treatment Groupsa Figure 1 shows the results for cardiovascular and non- Active cardiovascular mortality without inclusion of on- Placebo Group Treatment Group treatment systolic BP in the regression models. On- Variable (n = 2225) (n = 2358) treatment diastolic BP did not affect cardiovascular mortality, Patient-years of follow-up 5219 14 511 either during active treatment or during placebo adminis- Cerebrovascular eventsb 93 (18.3) 151 (10.7) tration. In contrast, noncardiovascular mortality was sig- c Coronary heart disease events 71 (13.8) 181 (12.7) nificantly and curvilinearly associated with diastolic BP, both Cardiovascular eventsd 193 (39.0) 404 (29.4) Ͻ Ͻ Mortality with active treatment (P .005) and with placebo (P .05). Cardiovascular 66 (12.6) 174 (12.0) Low on-treatment diastolic BP was associated with higher Noncardiovascular 62 (11.9) 161 (11.1) noncardiovascular mortality. For example, the HR for non- Cancer 27 (5.2) 59 (4.1) cardiovascular mortality for 5–mm Hg lower diastolic BP, Unknown cause 1 (0.2) 9 (0.6) from 65 to 60 mm Hg, was 1.15 (95% confidence interval [CI], 1.00-1.31) with active therapy and 1.28 (95% CI, 1.03- a Data are given as number of events (rate per 1000 patient-years) unless 1.59) with placebo. Results were similar for cancer mor- otherwise indicated. b Includes stroke and transient ischemic attack. tality, for which the HRs were, respectively, 1.26 (95% CI, c Includes sudden death, myocardial infarction, and coronary 1.01-1.58) and 1.36 (95% CI, 0.98-1.88). revascularization. Figure 2 shows the results for cerebrovascular events d Includes cardiovascular death, myocardial infarction, coronary revascularization, stroke, transient ischemic attack, heart failure, and aortic and CHD events. In the placebo group, the association aneurysm. between on-treatment diastolic BP and the risk of a cerebrovascular event was curvilinear, indicating an increased risk at low diastolic BP levels (PϽ.01); for ex- clusion in the trial, 29.9% had previous cardiovascular ample, the HR for 5–mm Hg lower diastolic BP, from 65 complications, and 14.5% had a history of CHD. to 60 mm Hg, was 1.25 (95% CI, 1.01-1.54). This was not the case with active treatment. The CHD risk re- BP AND EVENTS DURING FOLLOW-UP mained unchanged at lower diastolic BP with both active therapy and placebo. No significant association was In the placebo group, 52 patients (2.3%) were lost to fol- found between the risk of all cardiovascular events and low-up by the end of phase 1; after longer follow-up, 284 on-treatment diastolic BP during either active treatment patients (12.0%) in the active treatment group were lost or placebo administration (data not shown). to follow-up by the end of phase 2. Lost to follow-up was Figure 3 shows the relationship between on- defined as no information available for more than 1 year. treatment diastolic BP and the aggregate of all cardio- At the last follow-up visit, 2070 patients in the active treat- vascular events according to the history of CHD at base- ment group and 1782 patients in the placebo group re- line. During active treatment, there were 25.6 ceived active study medication or placebo tablets; the study cardiovascular events per 1000 patient-years in the 2022 medication was combined with open-label antihyperten- patients without CHD at baseline and 56.7 events per 1000 sive drugs in 220 actively treated patients and 51 patients patient-years in the 336 patients with CHD. In the pla- in the placebo group. One hundred fifty-seven patients in cebo group, there were 24.6 events per 1000 patient-

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1886

Downloaded From: https://jamanetwork.com/ on 09/30/2021 A Cardiovascular B Noncardiovascular 1.6 Active treatment Active treatment

1.4

1.2

1.0 Hazard Ratio

0.8

P = .44 P<.005 0.6

C D

1.6 Placebo Placebo

1.4

1.2

1.0 Hazard Ratio

0.8

P = .70 P<.05 0.6

95 85 75 65 55 95 85 75 65 55 Diastolic BP, mm Hg Diastolic BP, mm Hg

Figure 1. Hazard ratios for cardiovascular (A and C) and noncardiovascular mortality (B and D) in relation to on-treatment diastolic blood pressure (BP) (defined as low BP while receiving antihypertensive treatment) in the active treatment and placebo groups, adjusted for age, sex, smoking status, and previous antihypertensive treatment at baseline, and body weight, stroke, myocardial infarction, or diabetes at baseline and during follow-up. The hazard ratio gives the risk associated with a 5–mm Hg lower diastolic BP.

years in the 1897 patients without CHD and 73.3 events line and on-treatment diastolic BP was not significant per 1000 patient-years in the 328 patients with CHD. Low (P=.13). Results were similar when on-treatment sys- diastolic BP during active treatment was not associated tolic BP was included as a continuous time-dependent with increased risk in patients free of CHD at baseline, covariate in the analyses (data not shown). whereas the risk increased when on-treatment diastolic BP was lower than about 75 mm Hg in patients with CHD PROGNOSTIC SIGNIFICANCE (PϽ.02), with an HR of about 1.1 at 70 mm Hg. How- OF DIASTOLIC BP: ANALYSIS BY 5–mm Hg ever, the interaction between CHD status at baseline and CUTOFF CATEGORIES on-treatment diastolic BP was not significant (P=.19). In the placebo group, there was no support for an influ- All analyses were repeated using 5–mm Hg cutoff cat- ence of low diastolic BP on the incidence of cardiovas- egories for on-treatment diastolic BP, as in the SHEP trial.15 cular events in patients with CHD at baseline, but low These analyses yielded results similar to those when dia- diastolic BP was associated with increased risk of car- stolic BP was used as a continuous variable. Similar to diovascular events in patients without CHD (PϽ.01). Figure 4 in Somes et al,15 our Figure 4 shows the re- However, the interaction between CHD status at base- sults for the aggregate of all cardiovascular events, with

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1887

Downloaded From: https://jamanetwork.com/ on 09/30/2021 A Cerebrovascular Events B CHD Events 1.6 Active treatment Active treatment

1.4

1.2

1.0 Hazard Ratio

0.8

P = .36 P = .49 0.6

C D

1.6 Placebo Placebo

1.4

1.2

1.0 Hazard Ratio

0.8

P<.01 P = .73 0.6

95 85 75 65 55 95 85 75 65 55 Diastolic BP, mm Hg Diastolic BP, mm Hg

Figure 2. Hazard ratios for cerebrovascular events (A and C) and coronary heart disease (CHD) events (B and D) in relation to on-treatment diastolic blood pressure (BP) (defined as low BP while receiving antihypertensive treatment) in the active treatment and placebo groups, adjusted for age, sex, smoking status, and previous antihypertensive treatment at baseline, and body weight, stroke, myocardial infarction, or diabetes at baseline and during follow-up. The hazard ratio gives the risk associated with a 5–mm Hg lower diastolic BP.

inclusion of on-treatment systolic BP as a dichotomized BP is associated with higher noncardiovascular mortal- variable in the model. There was no consistent associa- ity. In patients with evidence of CHD at baseline, low on- tion between on-treatment diastolic BP and the ad- treatment diastolic BP is associated with increased risk justed HR for these events with either active treatment of cardiovascular events, which is not the case in pa- or placebo. Results were similar when on-treatment sys- tients without a history of CHD. Inclusion of on- tolic BP was used as a continuous variable or was re- treatment systolic BP in the regression models did not moved from the models (data not shown). affect the relationships between on-treatment diastolic BP and outcome, which indicates that the findings were COMMENT not confounded by the achieved systolic BP. Many observational cohort studies reported on the The major findings of the present study in older pa- shape of the relationship between BP and risk in various tients with isolated systolic hypertension are that low- populations, including patients with hypertension re- ering of diastolic BP with active antihypertensive treat- ceiving BP-lowering treatment.1 The results have not al- ment, to as low as about 55 mm Hg, does not seem to ways been consistent. In addition, data from untreated increase cardiovascular mortality, whereas low diastolic subjects are unlikely to be applicable to treated hyper-

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1888

Downloaded From: https://jamanetwork.com/ on 09/30/2021 A No CHD at Baseline B CHD at Baseline 1.6 Active treatment Active treatment

1.4

1.2

1.0 Hazard Ratio

0.8

P = .84 P<.02 0.6

C D

1.6 Placebo Placebo

1.4

1.2

1.0 Hazard Ratio

0.8

P<.01 P = .13 0.6

95 85 75 65 55 95 85 75 65 55 Diastolic BP, mm Hg Diastolic BP, mm Hg

Figure 3. Hazard ratios for cardiovascular events in relation to on-treatment diastolic blood pressure (BP) (defined as low BP while receiving antihypertensive treatment) in patients without (A and C) and with (B and D) coronary heart disease (CHD) at baseline in the active treatment group and in the placebo group, adjusted for age, sex, smoking status, and previous antihypertensive treatment at baseline, and body weight, stroke, myocardial infarction, or diabetes at baseline and during follow-up. The hazard ratio gives the risk associated with a 5–mm Hg lower diastolic BP.

tensive patients, and observational studies on treated hy- priate for assessment of the existence of a J-curve for on- pertensive patients are difficult to interpret in the ab- treatment BP because results in treated patients can be sence of a control group. A recent secondary analysis of compared with those in a control group. In a meta- the International Verapamil-Trandolapril Study empha- analysis of individual patient data from 7 randomized clini- sized the existence of a J-curve in actively treated pa- cal trials involving 40 233 persons with primarily systolic- tients with hypertension with stable CHD.16 The find- diastolic hypertension,4-10 a J-curve relationship was ing of a J-curve may be the result of confounding factors observed between diastolic BP and the risk of cardiovas- and ill health, associated with low BP and high risk for cular and noncardiovascular mortality in both treated and events (reverse causation). This has been confirmed in untreated patients.3 The authors concluded that the in- a large Finnish observational study in treated hyperten- creased risk observed in patients with low BP was not sive patients,19 in whom the observed J-curve relation- related to antihypertensive treatment and could prob- ship between diastolic BP and cardiovascular and non- ably be explained by poor health conditions. This con- cardiovascular mortality could be ascribed to preexisting clusion is supported by the results from the European cardiovascular disease, resulting in low diastolic BP. Pla- Working Party on High Blood Pressure in the Elderly trial cebo- or nontreatment-controlled trials are more appro- in which patients with low on-treatment BP had lower

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1889

Downloaded From: https://jamanetwork.com/ on 09/30/2021 A B

5 Active treatment Placebo

2 Hazard Ratio

90 85 80 75 70 65 60 55 95 85 80 75 70 65 60 55 Diastolic BP Cutoff, mm Hg Diastolic BP Cutoff, mm Hg

Figure 4. Hazard ratios for cardiovascular events according to on-treatment diastolic blood pressure (BP) (defined as low BP while receiving antihypertensive treatment) by 5–mm Hg cutoff values in the active treatment (A) and placebo (B) groups, adjusted for age, sex, smoking status, and previous antihypertensive treatment at baseline; on-treatment systolic BP as a dichotomous variable; and body weight, stroke, myocardial infarction, or diabetes at baseline and during follow-up.

values for body mass index and hemoglobin, which can uretic (chlorthalidone), with the possible addition of the be considered indicators of poor health.20 ␤-blocker atenolol or reserpine for better BP control; in When older patients with isolated systolic hyperten- the Syst-Eur Trial, treatment was started with the dihy- sion are treated with antihypertensive drugs, not only sys- dropyridine calcium channel blocker nitrendipine, to tolic but also diastolic BPs are reduced.8,12,17,21 It has been which the converting enzyme inhibitor enalapril and hy- argued that patients with isolated systolic hypertension drochlorothiazide could be added. Patients with CHD at and, thus, by definition having low diastolic BP, would baseline were not analyzed separately in the SHEP trail. be particularly vulnerable to the consequences of treat- About 5% of the randomized patients had a history of myo- ment-induced further lowering of diastolic BP.1 This is cardial infarction8; thus, it is unlikely that the percent- an important issue of clinical relevance because dia- age of patients with CHD exceeded the prevalence of CHD stolic BP may reach very low values in these patients be- at baseline in the Syst-Eur Trial. fore systolic BP is normalized. The SHEP investigators In contrast to the findings on cardiovascular mortal- concluded that lower achieved diastolic BP was associ- ity, we observed increases in noncardiovascular and can- ated with increased risk of CHD, stroke, and cardiovas- cer mortality at low diastolic BP, both with active treat- cular disease in actively treated patients; thus, some pa- ment and with placebo. These findings are compatible tients may have been overtreated.15 The relative risk for with the notion that the J-curve may be explained by ill the combined cardiovascular events became significant health and reverse causation. These data could also sug- for diastolic BP less than 70 mm Hg and approached a gest that one should be careful in treating patients at high 2-fold increase in risk for diastolic BP less than 55 mm Hg. risk of noncardiovascular death. This adverse relationship was not observed in the An unexpected observation in the placebo group was control group. Our overall results differ from those of that low diastolic BP was associated with a higher inci- the SHEP trial. In the active treatment group, we have dence of cerebrovascular events and not of CHD. How- not observed significant relationships between on- ever, the former finding agrees with the pooled analysis treatment diastolic BP and, respectively, cardiovascular of 9 epidemiologic studies in patients with isolated sys- mortality, cerebrovascular events, CHD events, and an tolic hypertension in which an inverse relationship be- aggregate of cardiovascular events, also when applying tween stroke mortality and diastolic BP was observed.22 the same statistical methods as in the SHEP trial.15 That lower diastolic BP was not associated with in- Several differences between the SHEP trail and the creased risk of cerebrovascular events during active treat- Syst-Eur Trial should be considered. The cutoff value for ment in the Syst-Eur Trial is compatible with a greater diastolic BP required for randomization into the trial was reduction in stroke risk23 and a greater decrease in ca- 90 mm Hg in the SHEP trial and 95 mm Hg in the Syst- rotid intima-media thickness24 with regimens based on Eur Trial. The SHEP trial investigators could analyze on- calcium channel blockers than with other regimens. treatment diastolic BP down to 25 mm Hg, whereas dia- Some limitations of the present post hoc analysis must stolic BP was rarely less than 55 mm Hg in the Syst-Eur be considered. All patients received active study treat- Trial. First-line treatment in the SHEP trial was a di- ment after the end of the double-blind phase of the trial,

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1890

Downloaded From: https://jamanetwork.com/ on 09/30/2021 which increased the number of observations in the ac- tion of the BIOMED Research Program sponsored by the tive treatment group. The number of observations and, European Union. consequently, the power of the statistical analyses are smaller in the placebo group, which explains the larger REFERENCES CIs. The risk of low diastolic BP could be greatest for CHD events, particularly in patients with evidence of CHD at 1. Cruickshank J. The J-curve in hypertension. Curr Cardiol Rep. 2003;5(6):441-452. baseline. Whereas we observed an increased risk for an 2. Hansson L, Zanchetti A, Carruthers SG, et al; HOT Study Group. Effects of in- tensive blood-pressure lowering and low-dose aspirin in patients with hyperten- aggregate of cardiovascular events at low diastolic BP in sion: principal results of the Hypertension Optimal Treatment (HOT) ran- patients with evidence of CHD at baseline, a similar but domised trial. Lancet. 1998;351(9118):1755-1762. 3. Boutitie F, Gueyffier F, Pocock S, Fagard R, Boissel JP; INDANA (Individual Data nonsignificant trend was observed for incident CHD (data Analysis of Antihypertensive Intervention) Project Steering Committee. not shown), related to the small number of events and J-shaped relationship between blood pressure and mortality in hypertensive pa- the limited power of the analysis. tients: new insights from a meta-analysis of individual-patient data. Ann Intern Med. 2002;136(6):438-448. In conclusion, these findings support the hypothesis that 4. Hypertension Detection and Follow-up Program Cooperative Group. Five-year find- antihypertensive treatment can be intensified for the pre- ings of the hypertension detection and follow-up program, I: reduction in mor- vention of cardiovascular events when systolic BP is not un- tality of persons with high blood pressure, including mild hypertension. JAMA. 1979;242(23):2562-2571. der control in older patients with isolated systolic hyper- 5. Amery A, Birkenhager W, Brixko P, et al. Mortality and morbidity results from tension, at least until diastolic BP reaches about 55 mm Hg. the European Working Party on High Blood Pressure in the Elderly trial (EWPHE). Lancet. 1985;1(8442):1349-1354. However, a prudent approach is warranted in patients with 6. Medical Research Council Working Party. MRC trial of treatment of mild hyper- isolated systolic hypertension and concomitant CHD, in tension: principal results. BMJ. 1985;291:97-104. whom diastolic BP should probably not be lowered to less 7. Coope J, Warrender TS. Randomised trial of treatment of hypertension in elderly patients in primary care. BMJ. 1986;293(6555):1145-1151. than 70 mm Hg, when the relative risk reaches about 1.1. 8. SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of Accepted for Publication: May 20, 2007. the Systolic Hypertension in the Elderly Program (SHEP). JAMA. 1991;265 (24):3255-3264. Correspondence: Robert H. Fagard, MD, Hypertension 9. Da¨hlof B, Lindholm LH, Hansson L, Scherste´n B, Ekbom T, Wester PO. Morbid- and Cardiovascular Rehabilitation Unit, U Z Gasthuisberg- ity and mortality in the Swedish Trial in Old Patients with Hypertension (STOP-Hypertension). Lancet. 1991;338(8778):1281-1285. Hypertensie, Herestraat 49, B-3000 Leuven, Belgium 10. Medical Research Council Working Party. Medical research council trial of treat- ([email protected]). ment of hypertension in older adults: principal results. BMJ. 1992;304(6824): Author Contributions: Dr Fagard had full access to all of 405-412. 11. Wang JG, Staessen JA, Franklin SS, Fagard R, Gueyffier F. Systolic and diastolic the data in the study and takes responsibility for the integ- blood pressure lowering as determinants of cardiovascular outcome [published rity of the data and the accuracy of the study analysis. Study online ahead of print April 18, 2005]. Hypertension. 2005;45(5):907-913. concept and design: Fagard, Bulpitt, and de Leeuw. Acqui- doi:10.116/01.HYP.0000165020. 12. Staessen JA, Fagard R, Thijs L, et al. Randomised double-blind comparison of sition of data: Fagard, Staessen, Celis, Bulpitt, de Leeuw, placebo and active treatment for older patients with isolated systolic hyperten- Tuomilehto, and Yodfat. Analysis and interpretation of data: sion. The Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Lancet. 1997;350(9080):757-764. Fagard, Staessen, Thijs, Bulpitt, de Leeuw, and Leonetti. 13. The Australian therapeutic trial in mild hypertension: report by the Management Drafting of the manuscript: Fagard, de Leeuw, and Yodfat. Committee. Lancet. 1980;1(8181):1261-1267. Critical revision of the manuscript for important intellectual 14. Liu L, Wang JG, Gong L, Liu G, Staessen JA. Comparison of active treatment and placebo in older patients with isolated systolic hypertension. J Hypertens. content: Staessen, Thijs, Celis, Bulpitt, de Leeuw, Leonetti, 1998;16(12, pt 1):1823-1829. and Tuomilehto. Statistical analysis: Fagard, Thijs, and 15. Somes GW, Pahor M, Shorr RI, Cushman WC, Applegate WB. The role of dia- Bulpitt. Obtained funding: Fagard, Staessen, Bulpitt, and stolic blood pressure when treating isolated systolic hypertension. Arch Intern Med. 1999;159(17):2004-2009. Tuomilehto. Administrative, technical, and material sup- 16. Messerli FH, Mancia G, Conti R, et al. Dogma disputed: can aggressively lower- port: Staessen, Celis, and de Leeuw. Study supervision: ing blood pressure in hypertensive patients with coronary artery disease be dangerous? Ann Intern Med. 2006;144(12):884-893. Fagard, Staessen, Celis, and de Leeuw. 17. Staessen JA, Thijs L, Fagard R, et al. Effects of immediate versus delayed anti- Financial Disclosure: None reported. hypertensive therapy on outcome in the Systolic Hypertension in Europe Trial. Funding/Support: This study was supported by Bayer AG J Hypertens. 2004;22(4):847-857. 18. Allison PD. Estimating Cox Regression Models with PROC PHREG. Survival Analy- (Wuppertal, Germany) and the National Fund for Sci- sis Using the SAS System: A practical guide. Cary, NC: SAS Institute Inc; 1995: entific Research (Brussels, Belgium). Study medication 111-184. was donated by Bayer AG and Merck & Co Inc (White- 19. Tuomilehto J, Ryyna¨nen OP, Koistinen A, Rastenyte D, Nissinen A, Puska P. Low diastolic blood pressure and mortality in a population-based cohort of 16,913 house Station, New Jersey). The trial was carried out in hypertensive patients in North Karelia, Finland. J Hypertens. 1998;16(9):1235- consultation with the World Health Organization, the In- 1242. 20. Staessen J, Bulpitt C, Clement D, et al. Relation between mortality and treated ternational Society of Hypertension, the European Soci- blood pressure in elderly patients with hypertension: report of the European Work- ety of Hypertension, and the World Hypertension League. ing Party on High Blood Pressure in the Elderly. BMJ. 1989;298(6687):1552- Role of the Sponsors: Investigators at the University of 1556. 21. Fagard RH, Van den Enden M. Treatment and blood pressure control in isolated Leuven initiated and designed the Syst-Eur Trial before systolic hypertension vs diastolic hypertension in primary care. J Hum Hypertens. seeking sponsorship. The funders had no role in the de- 2003;17(10):681-687. sign and conduct of the trial; collection, management, 22. Paultre F, Mosca L. Association of blood pressure indices and stroke mortality in isolated systolic hypertension [published online ahead of print May 5, 2005]. analysis, and interpretation of the data; or preparation, Stroke. 2005;36(6):1288-1290. doi:10.1161/01.STR.0000166201.79222.4d. review, or approval of the manuscript. 23. Turnbull F; Blood Pressure Lowering Treatment Trialists’ Collaboration. Effects of different blood-pressure-lowering regimens on major cardiovascular events: Previous Presentation: This study was presented at the results of prospectively-designed overviews of randomised trials. Lancet. 2003; 15th European Meeting on Hypertension; June 18, 2005; 362(9395):1527-1535. Milan, Italy. 24. Wang JG, Staessen JA, Li Y, et al. Carotid intima-media thickness and antihypertensive treatment: a meta-analysis of randomized controlled trials [pub- Additional Contributions: Nicole Ausseloos provided sec- lished online ahead of print June 8, 2006]. Stroke. 2006;37(7):1933-1940. doi: retarial assistance. The Syst-Eur trial was a concerted ac- 10.1161/01.STR.0000227223.90239.13.

(REPRINTED) ARCH INTERN MED/ VOL 167 (NO. 17), SEP 24, 2007 WWW.ARCHINTERNMED.COM 1891

Downloaded From: https://jamanetwork.com/ on 09/30/2021