Isolated Systolic Hypertension: New Answers, More Questions

Patients with isolated systolic hypertension (ISH) pose a common and often difficult management problem for practicing physicians. There is good evidence that ISH is associated with an increased risk for stroke and cardiovascular morbid events.1 Indeed, there are abundant data demonstrating that it is the systolic component of arterial pressure elevation that correlates best with morbidity and mortality attributable to hypertension.2 The common focus on diastolic pressure is due to its use as a marker in clinical trials of antihypertensive agents. For many years, however, it has been quite reasonable to ask whether attempts at treatment of ISH are worth the effort in terms of potential benefit. Important questions were: Will ISH respond to therapy? Is treatment of ISH safe? Is there any benefit from treatment? Which (if any) drugs are preferable? We now have at least partial answers to these questions. A commonly used definition of ISH is systolic blood pressure ≥160 and diastolic blood pressure <90 mm Hg.3 However, the reader must be careful to ascertain the definition used in published papers because other definitions have been used in the past. The prevalence of ISH is quite low until after age 50 when it increases rapidly. About one-fourth of women over 80 years old will have ISH compared to about one-fifth of men.4 The problem is complicated by the lability of blood pressure in patients with ISH. Colandrea et al5 demonstrated that the prevalence figures could be quite evanescent if based on only one reading. His initial screen of a sample of people in a retirement village showed an ISH prevalence of 13.9%, but this fell to 2.7% if an additional two blood pressure determinations were reviewed. Inaccurate indirect blood pressure readings also pose a problem; pseudohypertension exists to an uncertain degree in this population. Pseudohypertension is due to medial calcification of middle-size arteries (Mönckeberg's medial sclerosis) rather than occlusive atherosclerotic lesions.6 This entity places patients at risk for treatment of a disorder (hypertension) that they may not have. The mechanisms for disproportional increase of systolic hypertension have been fairly well elaborated.7 There is an increase in arterial wall collagen and loss of elastin, thereby resulting in a decrease of arterial compliance independent of arteriosclerosis. The stiffening and calcification due to arteriosclerosis further decreases compliance. It is important to note that ISH is not simply a manifestation of arteriosclerosis but carries its own independent morbidity and mortality. The questions of whether ISH responds to treatment at all and whether treatment is safe have been answered in the affirmative. The majority of elderly patients have low to normal plasma renin profiles and respond very nicely to strict dietary sodium restriction (45 mmol sodium per day) or to diuretics. The relatively small subset of patients with high plasma renin profiles responds less well or not at all to these measures.8 We do not recommend plasma renin profiling as a standard diagnostic tool, however. A therapeutic trial will yield practical results. The Systolic Hypertension in the Elderly Program (SHEP) pilot trial9 demonstrated that ISH could be successfully treated with small doses of chlorthalidone to which atenolol or reserpine could be added if necessary; this treatment proved to be remarkably safe. The Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents10 studied 51 patients with ISH and found that the response rate to 25–100 mg of hydrochlorothiazide was from 78% to 89% and that the response to one or two drugs was 94.1%. More frequent and severe hypokalemia occurred at the higher doses of diuretic. Indeed, 25 mg of hydrochlorothiazide reduced the blood pressure by 22.8/7.0 mm Hg at 2 weeks and was associated with a reduction of serum potassium of only 0.06 mmol/L. In addition, both the VA and SHEP trials performed psychometric testing and found no evidence of deterioration of cognitive function or mood.9–11 The recently published SHEP trial11 has provided us with the critical answer in regard to treatment of ISH: yes, there is benefit of drug therapy in reducing subsequent cardiovascular and cerebrovascular events. The SHEP investigators randomized 4,736 people with ISH that they had selected from a cohort of 447,921 screened. Their goal was 4,800 randomized patients; 98.7% of goal is a remarkable achievement for this type of study. Patients were randomly allocated to either an active treatment group or a placebo group with stratification by whether or not they were receiving antihypertensive drugs at the time of screening. This stratification later proved to have been unnecessary as the results were the same regardless of prior treatment. Entry criteria included age 60 years or above, systolic BP 160–219 mm Hg and diastolic BP less than 90 mm Hg. People with known cardiovascular and cerebrovascular disease were included as were diabetics. The blinded randomization process succeeded in producing two groups that were closely matched in terms of baseline characteristics. The average age was 71.6 years with 13.7% 80 years or older; 13.9% were black (but only 4.6% were black men); average BP was 170.3/76.6; and only 12.7% were current smokers. Indeed, 50.2% said that they had never smoked, a fact that may account in part for the low incidence rate of end-point events compared to that seen in the general population. Treatment was by a step-care algorithm in which patients were assigned to chlorthalidone 12.5 mg (or its placebo) and stepped up as needed to 25 mg chlorthalidone, atenolol 25–50 mg daily or reserpine .05–.1 mg daily. Drugs were titrated to a goal systolic BP of less than 160 mm Hg for patients whose baseline BP was 180 mm Hg or more or a decrease of at least 20 mm Hg for those patients whose baseline BP was 160 to 179 mm Hg. Over the course of the study, active treatment successfully reduced systolic BP by 11–14 mm Hg more than did placebo so that 65%–72% of patients achieved the defined goal BP. Placebo response rates were 32%–40%, in part because many of the placebo group patients eventually received treatment (see below). Adverse effect rates were generally higher in the treatment group, but the differences were neither striking nor statistically significant. Laboratory changes characteristic of thiazide diuretics were observed, but the serum potassium for the active treatment group was only 0.3 mmol/L less than that of the placebo group and only 1.0% of values were less than 3.2 mmol/L. The results were impressive. The total stroke incidence over 70 months was reduced by 36% (P = 0.0003), and the incidence of non-fatal myocardial infarction plus coronary death was 27% lower in the treatment group. This translated to 30 strokes and 55 major cardiovascular events prevented per 1,000 patients over 5 years. The protection extended across the range of age (young-old to old-old), gender, and race with the exception of black men of whom there were too few for statistical significance. Treatment with the diuretic did not cause an increase in sudden or rapid cardiac death. It is necessary to design large randomized trials such as SHEP so that the study criteria and data analyses are very rigid. Intention to treat analysis assesses the end results based on the category to which the patient was originally randomized irrespective of whether they remained in that category. Trials such as SHEP must also be humane and ethical. Patients randomized to active treatment who did not remain on treatment were still counted as being actively treated. Placebo patients whose BP exceeded safety limits required treatment, but they were still counted as placebo patients. In SHEP, about 10% of the active treatment patients did not stay on active treatment; 13% of the placebo patients were actually receiving treatment at the end of year one, 33% at year three, and 44% at year five. Thus, the above differences in stroke and cardiovascular end points were noted despite the narrowing of the actual differences in treatment status between the two groups. This suggests that SHEP may have underestimated the benefit of treatment of ISH. Two glaring questions remain unanswered. Given that a thiazide-like diuretic was associated with such excellent results in the SHEP trial, that thiazides have shown BP lowering efficacy in other studies of ISH,8,10,12 and that fear of increased deaths from thiazide-associated metabolic changes did not bear out, what is the evidence that supports the primary use of “newer” agents to treat ISH? There is none at present, and a major trial of this point may never be fiscally feasible. Given the modest number of actual events prevented per 1000 patient years, how does the cost-benefit equation change when more expensive drugs are substituted for thiazide diuretics? We do not pretend to know the answer to these questions and we may, indeed, never know. The results of SHEP and its predecessor trials appear to mandate treatment of patients with carefully documented ISH. It appears for now that the initial drug of choice should be a thiazide or thiazide-like diuretic.