Abstract
Background Patients with isolated systolic hypertension (ISH) have been noted to be less responsive to β-blockers than patients with essential hypertension (HTN). The purpose of this study was to determine the relationship between blood pressure (BP) and heart rate in ISH.
Methods A total of 619 patients underwent 24-hour ambulatory BP monitoring. Patients were grouped as normal, HTN, or ISH. Clinical characteristics, mean BPs, and mean heart rate were compared between the groups.
Results Two hundred seventy-one patients had normal BP, 98 had HTN, and 90 had ISH. Antihypertensives were used in 37% with normal BP, 51% with HTN, and 58% with ISH. The pulse pressure was highest for ISH (67 [10] mm Hg), followed by HTN (59 [12] mm Hg) and normal BP (49 [7] mm Hg; P < 0.0001). Heart rate in ISH was 71 [10] beats per minute; slower than that for normal BP (73 [11] beats per minute; P = 0.0464) and HTN (78 [12] beats per minute; P < 0.0001).
There was a positive relationship between diastolic BP and heart rate. In ISH, there was a negative relationship between systolic BP and heart rate (slope = −0.18; r = 0.24; P = 0.0209) and a positive relationship between diastolic BP and heart rate (slope = 0.19; r = 0.33; P = 0.0015).
Conclusions The noted relationship of heart rate to systolic BP and pulse pressure may in part explain the lower efficacy of β-blockers in patients with ISH. The rate-lowering effects of these medications may result in an increase in systolic and pulse pressures.
The clinical characteristics of isolated systolic hypertension (ISH) and essential hypertension (HTN) differ. Isolated systolic hypertension tends to occur in older people.1,2In general, it is more difficult to treat, often requiring multiple medications. Beta-blockers are also noted to have lower efficacy in ISH than they do in HTN.3–5The mechanism of ISH has been proposed to be related to increased stiffness of the large arteries, such as the aorta and large elastic arteries.6,7This increased stiffness results in increased vascular impedance, thereby increasing systolic blood pressure (BP) and decreasing diastolic BP.8In addition, reflected waves from peripheral vessels may increase the systolic BP.9The potential role of heart rate in the pathophysiology of ISH has not been explored. With diminishing heart rate, stroke volume increases to maintain cardiac output.10–12Increased stroke volume in the setting of stiff arteries may result in increased systolic BP. In addition, decreased heart rate is associated with a prolonged diastolic interval, which may lead to lower diastolic pressures. We therefore hypothesized that patients with ISH would have an inverse relationship between systolic BP and heart rate and a direct relationship between diastolic BP and heart rate.
MATERIALS AND METHODS
Data collection was approved by the Institutional Review Board of Northwestern Memorial Hospital. Between 1987 and 2000, a total of 619 patients underwent 24-hour ambulatory BP monitoring at Northwestern Memorial Hospital. From this database, we identified patients who met the following criteria: (1) technically adequate recording, (2) initial recording (follow-up recordings were not included in this analysis), (3) findings classified as either normal, HTN, or ISH. Normal BP (BP) was defined as a 24-hour mean systolic BP less than 140 mm Hg and 24-hour mean diastolic BP less than 90 mm Hg. Essential hypertension was defined as a 24-hour mean systolic BP ≥140 mm Hg and 24-hour mean diastolic BP ≥90 mm Hg. Isolated systolic hypertension was defined as a 24-hour mean systolic BP of 140 mm Hg or higher and 24-hour mean diastolic BP of lower than 90 mm Hg. Pulse pressure was calculated as the 24-hour mean systolic BP minus 24-hour mean diastolic BP.
Ambulatory BP was measured and recorded by using standard techniques. Blood pressure was measured oscillometrically, with algorithmic determination of systolic and diastolic pressures. Correlation of oscillometrically derived systolic and diastolic pressures with manual measurements were confirmed using a sphygmomanometer at the beginning and end of each 24-hour recording. All data were entered into a custom database. The following clinical data were collected with each 24-hour ambulatory BP acquisition: age, sex, height, weight, 24-hour mean systolic BP, 24-hour mean diastolic BP, 24-hour mean heart rate, and medications taken at the time of the recording.
Data Analysis
Differences among the 3 groups were assessed using analysis of variance (ANOVA). Post hoc pairwise comparisons were performed with an unpaired t test. Linear regression analysis was used to evaluate the relationship of BP parameters to heart rate. Multiple linear regression analysis evaluated the relationship of pulse pressure to the clinical parameters. P < 0.05 was considered statistically significant. Data analysis was performed with StatView 5.0.1 (SAS Institute Inc, Cary, NC). Data that follow are presented as mean (SD).
RESULTS
There were 271 patients with normal BP, 98 with HTN, and 90 with ISH. Table 1 demonstrates their clinical characteristics. As expected, the patients with ISH were older (61 [16] years) than those with HTN (54 [13] years; P = 0.0024 vs ISH) and normal BP (48 [15] years; P < 0.001 vs ISH). There was no difference in sex distribution nor height, weight, or body mass index. Antihypertensive agents were used in 37% of those with normal BP, 51% with HTN, and 58% with ISH. A mean of 1.6, 1.9, and 1.7 antihypertensive agents were used in the treated patients. Medications associated with lower heart rates including β-blockers, diltiazem, and verapamil were administered to 24% of those with normal BP, 23% of those with HTN, and 32% of those with ISH.
The systolic and diastolic BPs for each of the 3 groups were significantly different (P < 0.001 by ANOVA and for all post hoc pairwise comparisons). The patients with HTN had the highest values (systolic, 155 [12] mm Hg and diastolic, 96 [5] mm Hg), followed by those with ISH (systolic, 149 [8] mm Hg and diastolic, 83 [6] mm Hg). Those with normal BP, by definition, had normal values (systolic, 128 [7] mm Hg and diastolic, 79 [6] mm Hg). There were also significant differences (P < 0.0001 by ANOVA and for all post hoc pairwise comparisons) in pulse pressure among the 3 groups. However, the pulse pressure was highest for patients with ISH (67 [10] mm Hg), followed by those with HTN (59 [12] mm Hg) and those with normal BP (49 [7] mm Hg). Heart rate in the patients with ISH was 71 [10] beats per minute (bpm), which was significantly slower than the heart rate of those with normal BP (73 [11] bpm; P = 0.0464) and those with HTN (78 [12] bpm; P < 0.001).
Relationship of Blood Pressure to Heart Rate
The relationships of systolic and diastolic BPs to heart rate are shown in Figures 1–3. In patients with normal BP, there was no significant relationship between systolic BP and heart rate. There was a positive relationship between diastolic BP and heart rate (slope = 0.17; r = 0.29; P < 0.001). In the patients with HTN, there was no significant relationship between systolic BP and heart rate. There was a significant positive relationship between diastolic BP and heart rate (slope = 0.09; r = 0.22; P = 0.0297). Finally, in the patients with ISH, there was a significant negative relationship between systolic BP and heart rate (slope = −0.18; r = −0.24; P = 0.0209). There was a significant positive relationship between diastolic BP and heart rate (slope = 0.19; r = 0.33; P = 0.0015).
Figure 4 demonstrates the relationship of pulse pressure to heart rate in the patients with normal BP, HTN, and those with ISH. There was a significant negative relationship between pulse pressure and heart rate in the patients with normal BP (slope = −0.21; r = 0.34; P < 0.001) and ISH (slope = −0.36; r = 0.37; P < 0.001). There was no significant relationship between pulse pressure and heart rate in the patients with HTN.
Effect of Patient Characteristics
There was a significant negative relationship between age and heart rate in each group (normal BP: slope = −0.38; P < 0.001; HTN: slope = −0.34; P = 0.0022; ISH: slope = −0.52; P = 0.0015). Multiple linear regression analyses were performed to evaluate the relationship between pulse pressure and age, height, weight, and heart rate. The results are shown in Table 2. The multiple linear regression analyses show that only age and heart rate were significant predictors of pulse pressure in the patients with ISH. In the patients with HTN, only age was a significant predictor of pulse pressure. Finally, in those with normal BP, age, body mass index, and heart rate were all significant predictors.
Patients Not on Antihypertensives
One hundred seventy-one patients with normal BP, 48 patients with HTN, and 38 patients with ISH were not treated with antihypertensive agents. The clinical characteristics, BP, and heart rate values are shown in Table 3. The findings in this group were essentially the same as for the total group. Systolic and diastolic BPs for each group were significantly different (P < 0.0001 for all post hoc pairwise comparisons except systolic BP in HTN vs ISH, which was P = 0.01). Again, the patients with HTN had the highest values (systolic, 152 [10] mm Hg and diastolic, 96 [4] mm Hg), followed by those with ISH (systolic, 147 [6] mm Hg and diastolic, 84 [5] mm Hg), and those with normal BP (systolic, 128 [7] mm Hg and diastolic, 80 [6] mm Hg). Pulse pressure also differed between the groups (P < 0.0001 for all post hoc pairwise comparisons except HTN vs ISH, P = 0.002) and was highest for patients with ISH (63 [8] mm Hg), followed by patients with HTN (56 [10] mm Hg) and those with normal BP (49 [6] mm Hg). Mean heart rate did not differ among the 3 groups.
Relationship of Blood Pressure to Heart Rate in Patients Not on Antihypertensives
In patients with normal BP, there was no significant relationship between systolic BP and heart rate. There was a positive relationship between diastolic BP and heart rate (slope = 0.09; r = 0.16; P = 0.0388). In patients with HTN, there was no significant relationship between systolic or diastolic BP and heart rate. In distinction to patients on antihypertensives, those not on medications with ISH showed no significant relationship between systolic or diastolic BP and heart rate.
DISCUSSION
The present study shows that there is a universal positive relationship between diastolic BP and heart rate, consistent with the expected decline in pressure with longer diastolic intervals (slower heart rate). However, only patients with ISH being treated with antihypertensives demonstrated a relationship between systolic BP and heart rate. As hypothesized, this was an inverse relationship. This resulted in a significant negative relationship between pulse pressure and heart rate. Although this was noted in the patients with normal BP as well, it was not as steep. Although the relationship of BP and heart rate in the patients with ISH on antihypertensives does not establish causality, these findings support the hypothesis that bradycardia may be a contributing factor and could explain the lower efficacy of β-adrenergic blockers in patients with ISH.
Many authors have provided evidence that the mechanism behind ISH is a decrease in arterial compliance. Pasierski et al.6found that elderly patients with isolated systolic hypertension had significantly decreased arterial compliance compared with controls. In addition, the patients with ISH had a higher cardiac output. Similarly, Meaney et al.7found significant increase in aortic stiffness in elderly patients with ISH compared with both young normotensives and those older than 65 years. These authors also showed that aortic stiffness increases with age even in the normotensive group. A correlate of increased vascular stiffness is an increase in vascular impedance and pulse wave velocity, and an early return of pulse wave reflection in systole, which also contributes to the isolated increase in systolic BP.9
Given the pathophysiologic mechanisms described earlier, it has been suggested that not all antihypertensive medications are appropriate for the treatment of ISH in the elderly. Simon et al.4compared the effects of propranolol and sodium nitroprusside in patients with ISH. In patients younger than 35 years, arterial compliance and stroke volume were normal. This cohort responded to propranolol with a decrease in systolic BP. The older group (age >35 years) did not display systolic or diastolic changes with propranolol. This cohort displayed decreased arterial compliance at baseline, which decreased further with propranolol administration. In a randomized placebo-controlled trial of felodipine, metoprolol, and their combination for ISH in the elderly, metoprolol as monotherapy was not as effective as felodipine or both drugs in combination.13
In the current study, systolic BP was negatively correlated with heart rate in the patients with ISH, whereas there was no correlation in normotensives or those with HTN. In addition, pulse pressure was inversely related to heart rate in the normotensives and the patients with ISH but not those with HTN. These observations suggest that slower heart rates may contribute to the systolic BP elevations in ISH. Although not investigated in this retrospective study, others have shown that stroke volume increases with slower heart rates.10–12Nobrega et al.10demonstrated in patients with complete heart block and dual chamber pacemakers with a fixed pacing rate similar to resting heart rate (73 [3] bpm], that mean arterial pressures increased during static exercise owing to increased stroke volume (a combination of increased contractility and the Frank-Starling mechanism). This was opposed to normal heart rate increases in dual chamber pacing mode during which stroke volume was unchanged. In the previously mentioned trial of felodipine and metoprolol,13stroke volume was significantly higher in the metoprolol and combination therapy groups, whereas heart rate was significantly lower in both groups. Taken together, these data support the hypothesis generated by the current study.
Whereas the patients with normal BP not on antihypertensives had similar findings to the entire cohort, no association between heart rate and BP was seen in the patients with HTN and ISH. It is difficult to draw conclusions as to the mechanisms of this disparity, but the much smaller sample size of this subset of patients could play a role. Other than treatment differences, other parameters (age, heart rate, BP) were similar between the entire cohort and this subset.
LIMITATIONS
The main limitation to this study is that it is a retrospective analysis and does not confirm a causal role of heart rate in ISH. The disparate ages between the normal and ISH groups could in fact be the underlying etiology for both the lower diastolic BP and lower heart rates seen in the ISH group. Indeed, age was a significant predictor of pulse pressure in all patient groups. In addition, the 13-year span of this study encompasses a period during which an increase in antihypertensive medications occurred. Although this may have had an impact on the rates of successful treatment of HTN, it does not change the description by this study of a previously unrecognized relationship between heart rate and BP in patients with ISH. Another significant limitation is the use of antihypertensives with negative chronotropic effects. Other limitations include the use of oscillometric BP determination instead of intra-arterial, and the use of 24-hour mean values, which may obscure an overlap of BP measurements between groups. Finally, other clinical factors that may affect BP were not evaluated in this population.
CONCLUSIONS
As the pathophysiology of ISH differs from essential hypertension, it is not surprising that the hemodynamic effects related to heart rate may differ as do the response to medications. The noted relationship of heart rate to systolic BP and pulse pressure in patients with ISH may in part explain the lower efficacy of β-blockers in patients with ISH. The rate-lowering effects of these medications may result in an increase in systolic BP and pulse pressure. Whereas the present study demonstrated only associations, further efforts to evaluate whether there is a causal relationship between heart rate and systolic BP are warranted.