Attempted Forced Titration of Blood Pressure to <130/85 mm Hg in Type 2 Diabetic Hypertensive Patients in Clinical Practice: The Diastolic Cost

Abstract

The authors assessed the practicality and results of forced titrating of blood pressure to <130/85 mm Hg based on guidelines of the sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure in the setting of a clinical practice in 257 diabetic, hypertensive patients. Goal diastolic pressure was achieved in 90% of the patients, but goal systolic pressure was achieved in only 33%. In 57% of the patients, the attained diastolic pressure was ≤70 mm Hg, and in 20% of the cohort diastolic pressure was reduced to <70 mm Hg (mean, 60±1 mm Hg). Patients with final diastolic pressure <70 mm Hg were older, had a higher prevalence of coronary artery disease, and higher initial systolic and pulse pressures compared with patients with final diastolic pressure of 71–85 mm Hg. Thus, attempted lowering of blood pressure to <130/85 mm Hg is associated with excessive lowering of diastolic pressure in a significant number of patients. Whether the benefits of tight systolic control out‐weigh the risks of excessive diastolic reduction requires further prospective assessment.

Aggressive lowering of blood pressure (BP) in diabetic patients appears justified, in view of the increased cardiovascular (CV) morbidity and mortality rates in diabetes. ¹ , ² , ³ , ⁴ Several large clinical trials have provided evidence that more intensive lowering of arterial pressure in diabetes is indeed both feasible ⁵ and effective ⁵ , ⁶ , ⁷ in reducing CV morbidity. For example, in the Hypertension Optimal Treatment (HOT) study, ⁶ the lowest rate of major CV events was observed in treated hypertensive diabetic patients in whom the goal diastolic BP (DBP) was ≤80mm Hg, with an achieved mean of 81 mm Hg. In the UK Prospective Diabetes Study (UKPDS), ⁷ the rate of stroke and death related to diabetes was lower in diabetic individuals whose BP was reduced to an average of 144/82 mm Hg compared with patients subjected to less tight BP control, whose achieved BP was an average of 154/87 mm Hg. Whereas these studies favor reducing BP in diabetes below the precepts of standard practice, the feasibility of attaining target levels of <130/85 mm Hg has hardly been studied. In particular, the practicality of vigorous lowering of systolic BP (SBP) in diabetes to <130 mm Hg remained to be shown, as studies have traditionally targeted DBP, and the reduction in SBP was generally a surrogate outcome. Only in the past few years has the treatment paradigm shifted to SBP. ⁸ This change in treatment approach reflects the results of several observational studies ⁹ , ¹⁰ , ¹¹ in which SBP has been found to be a stronger predictor of cerebrovascular and CV morbidity than DBP. A recent large meta‐analysis conducted by the World Health Organization and the International Society of Hypertension (Blood Pressure Lowering Treatment Trialists' Collaboration) ¹² showed, moreover, that the level of SBP achieved in clinical trials was the strongest determinant of how effectively stroke and CHD were reduced. Nevertheless, the implications for DBP of forced lowering of SBP to <130 mm Hg in diabetic hypertensive patients are largely unstudied. This latter issue may be of significance in light of recent data from the Framingham Study ¹³ indicating that in the middle aged and elderly, who make up the majority of the type 2 diabetic population, risk for coronary heart disease increases with lower DBP at any level of SBP >120 mm Hg. Further, in several studies, ¹⁴ , ¹⁵ low DBP in treated hypertensive patients was associated with increased risk for myocardial infarction. In one report, ¹⁴ the relative risk for myocardial infarction at an achieved DBP of 60 mm Hg was two‐fold greater than that at a DBP of 100 mm Hg.

In the present study, we assessed the results of attempted lowering of BP to the target level of <130/85 mm Hg based on guidelines of the sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) for the treatment of hypertension in diabetes mellitus. Goal SBP levels were achieved in only one third of patients, and even this modest success rate was associated with a lowering of DBP to ≤70 mm Hg in more than half of the treated hypertensive diabetic patients.

PATIENTS AND METHODS

As of the last quarter of 1997, the JNC VI guidelines for the treatment of hypertension in diabetes have been adopted as the “standard care” at the Institute of Endocrinology, Metabolism and Hypertension, Tel Aviv‐Sourasky Medical Center. Implementation of these recommendations was reinforced through a series of departmental seminars, regular discussions, and case reviews. The basic treatment principles were as follows: 1) no initial placebo period or withdrawal of drugs; 2) attempted lowering of SBP to <130 mm Hg and of DBP to <85 mm Hg, within the limits of patients' compliance and tolerability toward BP reduction (e.g., orthostatic hypotension) or the various drugs, as judged by the treating staff member; and 3) individualized tailoring of medications based on: a) the use of flexible drug combinations at low to submaximal doses to minimize side effects, with gradual increments in dose or number of medications at 4–12 week intervals; b) the use of angiotensin‐converting enzyme (ACE) inhibitors whenever possible; and c) the use of β blockers in individuals with documented previous coronary events. We systematically reviewed the charts of all diabetic hypertensive patients (N=257) who were actively treated and regularly followed for hypertension at our clinic and had a minimal follow‐up period of 12 months. Patients whose treatment in our clinic was initiated after the publication of the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) guidelines were not included in this analysis due to the change in BP‐lowering goals.

Assessment of BP

BP was measured by the treating physician using cuff sphygmomanometry with the patient in the sitting position, after 5 minutes of rest. The mean of three stable consecutive measurements (intermeasurements differences, <5 mm Hg) at 2‐minute intervals was recorded. The duration of follow‐up on active treatment (i.e., attempted lowering of BP to target levels of <130/85 mm Hg) was 12–25 months (mean, 20 months). We considered the BP recorded at our clinic before the introduction of additional medications or any dose adjustments as the initial BP. Initial BP was <130/85 mm Hg in 14 subjects (6.6%), all of whom already had received antihypertensive treatment. These patients were included in the analysis of the cohort. Final BP was defined as that observed when no further changes in antihypetensive therapy were undertaken due to the fulfillment of any of the following conditions: 1) stable BP in the target zone (<130/85 mm Hg); 2) DBP of 50 mm Hg, even if the SBP was still >130 mm Hg; or 3) stable BP which either appeared unresponsive to further therapy or at which the addition of more drugs or dose increments were considered unsafe (e.g., orthostasis) or led to unacceptable side effects.

Statistical Analyses

Treated patients were divided into categories according to the BP achieved following treatment. Analysis of variance and chi‐square tests were used to compare baseline clinical variables among these groups of patients. Pearson correlation was calculated between all continuous parameters. We used forward stepwise regression analysis to construct a predictive model for the change in DBP following treatment. This model allows simultaneous assessment of the independent relationship between each clinical variable considered and the outcome. All variables correlating with the outcome variable (change in DBP) in the univariate analysis at a significance level of <0.2 were introduced into this model. Statistical significance was set at the 0.05 level.

RESULTS

Basal anthropometric and clinical features of the 257 patients are summarized in Table I. The initial mean BP for the entire cohort was 159/86 mm Hg, reflecting the fact that the majority of patients were already receiving antihypertensive therapy at baseline (mean number of drugs, 1.2 per patient). At the time final BP was recorded, 60% of the patients were receiving ACE inhibitors; 49% were receiving calcium channel blockers; 30% were on diuretics; 26% on β blockers; 19% on α blockers; 20% on nitrates; and 8 % on angiotensin II receptor antagonists. Following treatment, goal DBP was achieved in 90% of patients (232/257). Goal SBP of <130 mm Hg, however, was reached in only 33.5% of this cohort (86/257) and the desirable SBP and DBP was observed in just 32% of the treated subjects. Nevertheless, despite this apparent high failure rate with respect to an SBP goal, the mean achieved SBP in the entire cohort was 132±1 mm Hg. In approximately only 8% of patients did both SBP and DBP remain above the designated target zone.

Table I.

Basal Anthropometric and Clinical Features of 257 Hypertensive Type 2 Diabetic Patients

Age (yr)	65.3±0.6
Women/men	135/122
Body mass index (kg/m2)	29.2±0.5
Systolic BP (mm Hg)	159±2
Diastolic BP (mm Hg)	86±1
Pulse pressure (mm Hg)	57±1
Initial number of hypotensive drugs	1.2±0.1
Numeric parameters are presented as mean ± SEM. BP=blood pressure

We analyzed our patient population based on the therapeutic outcome. Table II presents the main clinical and anthropometric characteristics of patients achieving both SBP and DBP goals, patients achieving DBP or SBP goal only, and patients in whom both SBP and DBP remained above target levels. As shown, there were significant differences in the mean age, gender distribution, initial SBP (but not DBP), and pulse pressure among the four outcome groups. Patients with an achieved final BP of <130/85 mm Hg were somewhat younger (63±1.4 vs. 67.5±0.8 years; p<0.01) and had lower initial SBP (but not DBP) and pulse pressure compared with the patients with persistent systolic hypertension but controlled DBP. Additionally, relatively more women than men failed to achieve both SBP and DBP control (38.5% vs. 25.9% of the respective gender populations; p<0.01).

Table II.

Clinical and Anthropometric Features of Type 2 Diabetic Patients Based on Systolic and Diastolic Outcome

Systolic BP (mm Hg)	<130	≥130	<130	≥130
Diastolic BP (mm Hg)	<85	<85	≥85	≥85	p Value*
Number (%)	82 (32)	150 (58.3)	4 (1.6)	21 (8.2)
Age (yr)	63±1.4	67.5±0.8**	42±9	62.9±2.7	<0.001
Body mass index (kg/m2)	27.2±0.8	30.0±0.6	27±3.8**	29.2±0.5	<0.03
Women (%)	25.9	68.9	2.2	3	<0.03
Men (%)	38.5	46.7	10.8	3.9	<0.03
Initial systolic pressure (mm Hg)	147±3	165±2†	133±10	167±5†	<0.03
Initial diastolic pressure (mm Hg)	88±1	88±1	89±7	90.2±2.7	NS
Initial pulse pressure (mm Hg)	60±2	77±2†	45±7	77±5	<0.01
Numeric parameters are presented as mean ± SEM. BP=blood pressure; NS=nonsignificant; *analysis of variance; **p<0.01 vs. <130/85 mm Hg; † p<0.001 vs. <130/85 mm Hg

While DBP was controlled in most patients, attempts to reach target levels of both SBP and DBP resulted in lowering of DBP to ≤70 mm Hg in 146/257 (57%) of our patients, and to <70 mm Hg in about one fifth (53/257) of the cohort. Table III summarizes the clinical characteristics of the patients classified by their final DBP: <70 mm Hg; 70 mm Hg; 71–84 mm Hg, and ≥85 mm Hg. The large number of patients achieving DBP of 70 mm Hg likely resulted from a digit preference in the measurement process. The lowest DBP group had an attained DBP mean of 60±1 mm Hg. The mean DBP for the 146 patients with final DBP of ≤70 was 66±0.5 mm Hg. Among the patients with final DBP within the target range (<85 mm Hg), age, initial pulse pressure, and initial SBP were the highest in subjects with the lowest final DBP (<70 mm Hg). Patients with the lowest achieved DBP (<70 mm Hg) had the highest prevalence of preexisting ischemic heart disease (Figure 1A). The use of nitrates (but not of other drug classes) was also more common in this group than in patients with higher final DBP (Figure 1B). Finally, although the initial DBP was similar in all groups (II, III) the decline in DBP was the largest (approximately 25 mm Hg) in patients with the lowest final DBP (Figure 2). There were also overall negative correlations between final DBP and age (r=−0.249; p<0.01), initial pulse pressure (r=−0.179; p<0.01) and final pulse pressure (r=−036; p<0.001). Final DBP was also positively related to body mass index (r=0.241; p<0.01).

Table III.

Clinical Features of 257 Treated Hypertensive Type 2 Diabetic Subjects Based on Attained Final Diastolic Blood Pressure

Diastolic BP (mm Hg)	<70	70	71–84	≥85	p Value*
Number (%)	53 (21)	93 (36)	97 (38)	14 (5)
Age (yr)	70±1	67±1	61±1**	64±3	<0.0002
Body mass index (kg/m2)	27.3±1.4	28.6±0.6	30.3±0.8	30.6±1	NS
Initial pressures (mm Hg)
Systolic	166±4†	159±3	153±2	169±4	<0.05
Diastolic	86±2	89±1	89±1	88±2	NS
Pulse	80±3††	71±2	64±2	81±5	<0.003
Final pressures (mm Hg)
Systolic	128±2	131±2	133±1	149±2‡	<0.03
Diastolic	60±1‡‡	70‡‡	80±1‡‡	92±1‡‡	<0.0002
Number of hypotensive drugs	2.5±0.2	2.2±0.1	2.3±0.3	2.3±0.1	NS
Numeric parameters are presented as mean ± SEM. BP=blood pressure; NS=nonsignificant; *analysis of variance; **p<0.01 vs. ≤70 mm Hg; † p<0.04 vs. 71–84 mm Hg; †† p<0.002 vs. 71–84 mm Hg; ‡ p<0.04 vs. <70 mm Hg; ‡‡ p<0.001 for all among‐group comparisons

Figure 1.

Prevalence of preexisting coronary artery disease (CAD) (A) and use of nitrates (B) in 257 type 2 diabetic hypertensive patients classified by their achieved diastolic blood pressure

Figure 2.

Reduction in diastolic blood pressure (DBP) during intensified hypotensive therapy classified by the achieved DBP

When the change in DBP was subjected to forward multiple stepwise regression analysis, 52% of the variation in the attained reduction in DBP could be explained by the combination of initial SBP (20%; β coefficient = 0.789; p<0.0001), age (18%; β coefficient = 0.857; p<0.0001), weight (2%; β coefficient = −0.1; p<0.013), preexisting ischemic heart disease (1%; β coefficient = 2.72; p<0.03), and the interaction between age and initial pulse pressure (11%; β coefficient = −0.01; p<0.0001). Notably, neither the attained DBP nor the actual decline in DBP was associated with the use of any specific antihypertensive drug class or a definable combination of drugs.

DISCUSSION

The present study represents feedback from a clinical practice to the recommendation to lower BP in patients with diabetes mellitus to <130/85 mm Hg. We are unaware of previous reports of systematic attempts to achieve the recommended target levels for both SBP and DBP in diabetics in the setting of a clinical practice. To date, formal studies of intensive therapy in hypertensive individuals with diabetes mellitus have targeted DBP; the SBP outcome was a surrogate parameter. ⁵ , ⁶ , ⁷ Indeed, even the most aggressive BP‐lowering protocols, such as the Appropriate Blood Pressure Control in Diabetes (ABCD) ⁵ and HOT ⁶ studies, which strived for DBP of <75 mm Hg and <80 mm Hg, respectively, achieved SBP in the vicinity of 135–140 mm Hg, clearly higher than the JNC VI target level. Additionally, since the present report is based on individualized application of JNC VI recommendations rather than on a rigid study protocol assessing a single drug, it likely reflects what can be accomplished in the “real world”—and at what cost.

One finding of our analysis is that in type 2 diabetes mellitus, the goal DBP is easier to meet than the SBP goal: 90% of the patients achieved the <85 mm Hg levels, but SBP was successfully lowered to <130 mm Hg in only one third of this cohort. These findings are consistent with large clinical trial results in patients over 60 years of age. DBP was also better controlled than SBP in diabetic subjects participating in the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). ¹⁶ In a “treatment to goal”‐oriented study, Singer and coworkers ¹⁷ treated a non‐selected cohort of 437 hypertensive patients, with the goal of achieving BP levels of <140/90 mm Hg. Target DBP was achieved in 86% of patients, whereas only 63% of patients achieved the target SBP, although it was 10 mm Hg higher than the JNC VI‐recommended value.

More importantly, however, our data indicate that substantial lowering of DBP beyond recommended DBP targets appears to be an inevitable outcome of attempts to approach the target SBP. Indeed, this finding is hardly surprising in light of the mean baseline BP of 159/86 mm Hg and the older age of the patients. As the initial excess in SBP is much larger than the excess in DBP, and in the absence of hypotensive agents acting exclusively on SBP, pushing SBP down to <130 mm Hg will inevitably drive DBP to low levels in some patients. That this phenomenon has not been reported previously in diabetes probably reflects the fact that ambitious, guideline‐based therapeutic targets in clinical studies in diabetes have been defined for DBP, and not for SBP. Although the rate of use of diuretics in our cohort was relatively low, sub analysis of the BP‐lowering effect showed no association with any particular drug class. Nevertheless, diuretics have been shown to effectively lower SBP more than DBP in older individuals with isolated systolic hypertension. ¹⁸

Several clinical features appear to be associated with substantial DBP lowering when tight BP control is attempted in diabetes. First, baseline SBP was highest in patients with the lowest final DBP, perhaps reflecting lack of aortic compliance. Second, initial pulse pressure was highest in that same group. Third, age was inversely related to the final DBP. Fourth, excessive lowering of DBP was unrelated to any specific drug class. Finally, the prevalence of preexisting ischemic heart disease was nearly two‐fold greater in patients with final DBP of <70 mm Hg compared with patients with final DBP of ≥70 mm Hg. Collectively, these data suggest that in a significant fraction of older diabetic individuals with high SBP and wide pulse pressure, lowering of SBP toward the desirable <130 mm Hg target based on current treatment guidelines may be impossible without excessive lowering of DBP.

Because a lower DBP may carry greater CV risk, ¹³ , ¹⁴ , ¹⁵ particularly in patients with existing coronary disease, ¹⁹ , ²⁰ the problem of therapeutically induced “diastolic hypotension” in diabetes poses a clinical challenge. The choice between leaving SBP above the desirable range and lowering DBP to approximately 60 mm Hg or less may not be unique for diabetics, and likely applies to many older hypertensive individuals with isolated systolic hypertension and/or wide pulse pressure. Because SBP levels as low as <130 mm Hg are presently called for in diabetes, however, but not in older hypertensive subjects at large, this dilemma is likely to be encountered more commonly in diabetic individuals. Concerns that low DBP is particularly hazardous to patients with coronary artery disease ¹⁹ , ²⁰ cast a further burden on practicing clinicians attempting to implement current guidelines; not only is the prevalence of coronary disease in diabetes (including unrecognized pathology) as high as 50%, ²¹ , ²² but in the present study the prevalence of patients with known ischemic heart disease was the highest in subjects with low achieved DBP. A recent meta‐analysis ²² of several large‐scale clinical trials in hypertension provides reassuring information for lowering BP beyond the limits of current practice, in that SBP lowering was beneficial even when DBP was lowered to <70 mm Hg. Nevertheless, even this study concluded that benefit increased with a lower ratio of DBP to SBP lowering. One subanalysis of the SHEP study, ²³ moreover, did suggest that in treated subjects with isolated systolic hypertension, a decrease of 5 mm Hg in DBP significantly increased the adjusted risk for stroke, coronary heart disease, and CV events.

CONCLUSIONS

We do not dispute the recommendation to lower BP to <130/80 mm Hg in diabetic patients. We suggest, however, that until more information on the risk ratio of leaving SBP >130 mm Hg vs. lowering DBP to <60 mm Hg is available in diabetic patients, qualifications concerning the attained DBP be added to existing guidelines for target BP in diabetes, particularly in older subjects and/or in subjects with coronary artery disease. Whether DBP lowering in diabetes should be restricted to a final DBP of no less than 60 mm Hg remains subject to further discussion.