Prognostic value of systemic blood pressure response during exercise in a community-based patient population with hypertrophic cardiomyopathy

Patient selection and follow-up

The clinical features and long-term outcome of HCM in our large cohort of unselected, consecutively studied patients assessed longitudinally in a well defined regional population from the Tuscany region in central Italy have been recently described (14). Between 1983 and 1995, cycloergometer exercise testing was administered in our outpatient clinic as part of the standard evaluation for patients with HCM.

Of the 202 patients in this population, 76 were excluded from the study for the following reasons: severe exercise limitation due to congestive symptoms (New York Heart Association functional class III–IV; n = 19); refusal to comply with the protocol (n = 13); noncardiovascular physical disabilities incompatible with exercise (e.g., severe osteoarthrosis or neurologic deficits; n = 9); age >70 years (n = 8); prior documentation of life-threatening arrhythmias (ventricular fibrillation or sustained ventricular tachycardia; n = 5), and exercise testing performed at other institutions with different methodology and protocol (n = 22). Therefore, the final study group comprised 126 patients (Table 1).

Mean age at the time of the exercise test was 42 ± 14 years (range 19 to 68); 89 patients were ≤50 years old, and 37 patients were >50 years old. Ninety (71%) were male; 121 patients were in sinus rhythm, whereas the remaining five were in atrial fibrillation. Overall duration of follow-up from the initial diagnosis of HCM was 9.4 ± 6.5 years (range 1 to 29); the follow-up period after the exercise test was 4.7 ± 3.7 years (range 1 to 13).

The diagnosis of HCM in our study patients was based on echocardiographic identification of a hypertrophied and nondilated left ventricle in the absence of another cardiac or systemic disease capable of producing the magnitude of left ventricular hypertrophy evident in that patient (30). No patient identified as having HCM solely in the course of a systematic pedigree analysis was included in the present series (10,20,31); however, 18 patients were evaluated after another relative had become part of the study group. Initial clinical evaluation was taken as the time the diagnosis of HCM was first confirmed. The most recent clinical evaluation was obtained in October to December 1995.

Ages at initial evaluation ranged from 1 to 68 years (mean 36 ± 15); 15 (12%) were <20 years, and four (3%) were >60 years. Ages at the most recent evaluation ranged from 13 to 74 years (mean 47 ± 14). After initial identification, patients were followed in a standard fashion at about one-year intervals with clinical examination, two-dimensional echocardiogram, 12-lead electrocardiogram (ECG) and 24- or 48-h ambulatory (Holter) ECG.

Exercise testing

Patients had cardioactive medications withdrawn at least five half-lives before the exercise test, with the exception of the 12 patients receiving amiodarone; each fasted ≥4 h before the test. Maximum, symptom-limited exercise tests were performed on a bicycle ergometer (Quinton Q 2000, Seattle, Washington) in the upright position. The same protocol was used with all patients, with an initial workload of 25 W, stepwise workload increments of 25 W every 3 min and a recovery phase comprising 3 min against a workload of 25 W followed by another 3 min at 0 W. Three leads were monitored in a continuous fashion (D2, V2 and V5); in addition, a 12-lead ECG was recorded at baseline, and at each minute during exercise and the recovery phase. Blood pressure was measured using a mercury sphygmomanometer at baseline, at 60 and 150 s of each step, and every minute during the recovery phase. Maximum predicted heart rate and workload for each patient was calculated with respect to age, gender, height and weight (32). Percent of functional capacity was calculated as: (maximum workload performed/maximum predicted workload) × 100.

An abnormal blood pressure response to exercise was defined as follows: 1) exercise-induced hypotension; any decrease in systolic blood pressure below baseline values occurring during the exercise phase or recovery in the absence of an initial rise with exercise, or alternatively a sustained (>1 min in duration) decrease of ≥20 mm Hg during exercise after an initial rise, and 2) failure to increase blood pressure with exercise, defined as a systolic blood pressure rise of less than 20 mm Hg from baseline during exercise.

Echocardiography

Echocardiographic studies were performed using commercially available instruments (Toshiba 65A and 270) and 3.5- or 5.0-mHz transducers. Extent and distribution of left ventricular hypertrophy was assessed from the two-dimensional echocardiogram as previously described (33,34). Magnitude of the peak instantaneous left ventricular outflow tract gradient under basal conditions was estimated with continuous wave Doppler (35). Significant left ventricular outflow tract obstruction was considered present when the peak instantaneous gradient was ≥30 mm Hg.

Statistical analysis

Data were expressed as mean ± standard deviation. Statistical analyses were performed using unpaired Student ttest for the comparison of normally distributed data. Chi-square test was used to compare noncontinuous variables expressed as proportions. Univariate analyses for survival curves were calculated using Kaplan-Meier estimates (36). Univariate and multivariate regression analysis, for the identification of independent predictors of HCM-related death, were performed by the Cox regression model (37), using a SPSS statistical package (SPSS, Chicago, Illinois). The multivariate analysis used the forced entry method, and only included those demographic, clinical and echocardiographic variables that were significantly associated with HCM-related mortality at univariate analysis.

Exercise testing

The 126 study patients exercised for 12 ± 4 min (range 2 to 25), reaching maximum workloads varying from 25 to 225 W; percent of maximum predicted heart rate achieved was 87 ± 14%; mean percent of predicted functional capacity was 75 ± 23% (Table 1). Reasons for termination of the test were as follows: fatigue (67 patients; 53%), dyspnea (26 patients; 21%), chest pain (21 patients; 16%), symptomatic hypotension (7 patients; 6%) and arrhythmias (5 patients; 4%). In 104 patients, ≥75% of the predicted maximal heart rate value was achieved during exercise, whereas in the other 22 patients testing was terminated before such levels could be achieved. Complete exercise test results for the study group are shown in Table 1.

During the exercise test, 33 patients (26%) had arrhythmias, including 29 with multiple ventricular premature beats, 2 with paroxysmal atrial fibrillation and 1 each with nonsustained ventricular tachycardia or nonsustained supraventricular tachycardia. In four patients exercise testing elicited rate-dependent conduction defects (complete or incomplete left bundle branch block) that had been absent at rest, whereas 23 others had ST-segment and T-wave alterations, including 16 patients in whom ST-segment depression suggested myocardial ischemia (i.e., ≥2 mm).

Systemic blood pressure response to exercise

Of the 126 patients, 98 (78%) showed normal blood pressure increase during the exercise test (mean 60 ± 25 mm Hg; range 25 to 110) (Table 1). The remaining 28 (22%) had an ABPR, including 19 patients in whom blood pressure failed to increase appropriately (maximum increase 10 ± 6 mm Hg; range 5 to 18), and nine others with hypotension (maximum blood pressure drop 28 ± 6 mm Hg; range −20 to −35); six of these nine had an initial blood pressure rise followed by subsequent drop ≥20 mm Hg, whereas the other three had blood pressure drop below baseline values in the absence of an initial rise (Fig. 1; Table 2). Patients with ABPR more frequently experienced dizziness during the exercise test than patients with a normal blood pressure response (25% vs. 2%, p < 0.001; Table 1), but there were no other significant differences between the two groups with regard to the occurrence of symptoms and arrhythmias (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Diagram showing systolic blood pressure (BP) and heart rate responses during exercise and the recovery phase in a severely symptomatic 16-year old girl with nonobstructive hypertrophic cardiomyopathy (patient #1 in Table 2). Systolic blood pressure failed to increase from the baseline value of 100 mm Hg, and the test was interrupted after 2 min and 30 s due to hypotension (baseline systolic blood pressure of 100 mm Hg dropping abruptly to 80 mm Hg) associated with dizziness. This patient died suddenly and unexpectedly six months after the exercise test.

Compared with patients with normal blood pressure response, the 28 patients with ABPR achieved a lesser degree of cardiovascular stress as judged by maximum heart rate and double product values (Table 1). Patients with ABPR were also younger and more often female, had smaller left ventricular end-diastolic dimension and more often experienced chest pain unrelated to exercise testing. The patient subgroups with and without ABPR did not differ in terms of the frequency of drug treatment with amiodarone, beta-adrenergic blocking agents, calcium channel antagonists, angiotensin-converting enzyme inhibitors or diuretics during the follow-up period (Table 1).

Peak systolic left ventricular outflow tract gradient (≥30 mm Hg; range 30 to 110 mm Hg) measured by Doppler echocardiography under basal conditions before exercise testing was present in 26 of the 126 study patients. Outflow gradients ≥30 mm Hg were present with similar frequency in patients with ABPR (9/28; 32%) and those with normal pressure response to exercise (18/98; 18%; p = NS) (Table 1). Average outflow tract gradient was also similar in the two patient subgroups (28 ± 34 mm Hg with ABPR vs. 32 ± 26 mm Hg without ABPR; p = NS).

Cardiovascular mortality and predictive value of abnormal blood pressure response

During the follow-up period, nine patients died of HCM-related causes, at 17 to 69 years of age (mean 43 ± 18). Duration of time between the exercise test and death was 5.6 ± 3.1 years (range, 5 months to 8 years). Of the 9 HCM-related deaths, six were primarily heart failure–related and three were sudden and unexpected. Of the 6 patients who died of heart failure, 2 had ABPR (one with hypotension and one with failure to appropriately increase blood pressure). Three patients died suddenly, each before the age of 30; two of these patients had ABPR, manifested by symptomatic exercise-induced hypotension in both cases (Table 1; Fig. 1). Therefore, of the 9 patients with HCM-related death, a total of 4 (44%) had ABPR.

Of the other 117 surviving patients, a total of 24 (20%) had ABPR, of whom six had exercise-induced hypotension and 18 failed to appropriately increase blood pressure. The negative predictive accuracy of ABPR for cardiovascular mortality was 95% and positive predictive accuracy was 14%.

In the subset of patients aged 50 or less at the time of the exercise test (n = 89), Kaplan-Meier survival analysis showed an increased risk for HCM-related mortality in patients with ABPR as compared with those patients with a normal blood pressure response to exercise (p = 0.04; Fig. 2). Conversely, in those patients >50 years of age (n = 37), ABPR was not associated with an increased risk for cardiovascular mortality (p = 0.4). Only one of the nine HCM-related deaths occurred among patients >50 years of age (in a 68-year-old man with a normal blood pressure response during exercise), whereas the remaining eight deaths occurred among patients ≤50 years. The two subgroups of patients, aged >50 years and ≤50 years at the time of the exercise test, had a similar prevalence of ABPR (24% vs. 19%, respectively; p = NS), and comparable follow-up duration (4.5 ± 3.6 years vs. 3.7 ± 2.9 years, respectively; p = NS).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cumulative survival measured from the time of the maximal symptom-limited exercise test in 89 hypertrophic cardiomyopathy patients aged ≤50 years. The 21 patients who demonstrated abnormal systemic blood pressure response to exercise (ABPR; broken line) showed significantly reduced survival compared with the 68 patients with a normal blood pressure (BP) response (solid line), over an average follow-up period of 4.7 ± 3.7 years.

Univariate regression analysis showed only ABPR, age, history of syncope and atrial fibrillation, out of 20 demographic, clinical and echocardiographic features (Table 1), to be significantly associated with HCM-related mortality in patients ≤50 years of age. Of these four variables, the only independent predictors of risk at multivariate analysis were ABPR (p = 0.04; odds ratio = 4.5; 95% confidence interval: 1.1, 20.1) and atrial fibrillation (p = 0.03; odds ratio = 5.4; 95% confidence interval: 1.1, 7.5). For those patients ≤50 years of age, negative predictive accuracy of ABPR for cardiovascular mortality was 94% and positive predictive accuracy was 19%. The presence of nonsustained ventricular tachycardia on Holter ECG monitoring was not significantly associated with HCM-related mortality in our study population at univariate analysis (p = 0.4), and therefore was not included in the multivariate model.

Risk stratification in HCM

Sudden and unexpected death, the most devastating feature of the natural history of HCM, is a relatively uncommon event, but may constitute the first clinical manifestation of the disease in previously asymptomatic patients (1–7,18,19,22–26). Despite extensive efforts over three decades, the assessment of high (and low) risk in individual patients with HCM remains largely unresolved (6,7)due to the relatively low prevalence of the disease, extreme heterogeneity and complex pathophysiology characteristic of the disease, as well as the severe degree to which tertiary center patient referral bias has dominated the HCM literature and has skewed our perceptions of this condition (6,7,29). A hypotensive blood pressure response to exercise has been regarded as evidence for hemodynamic instability and proposed as a marker of increased risk for adverse cardiovascular events and mortality, both in coronary artery disease (38–41)and in selected tertiary center patient populations with HCM (24,27,28). For these reasons, we have investigated the clinical implications of the systemic blood pressure response to exercise in a large community-based HCM cohort in which the selection of patients was not biased by the preferential referral of those already judged to be at high risk.

Prevalence of ABPR

In our community-based outpatient population, exercise testing was administered as a part of a standard noninvasive evaluation of patients with HCM. Of the 126 patients in the study group, about 20% showed an ABPR that included the development of hypotension or the failure to appropriately increase blood pressure with exercise. The prevalence of ABPR in the present study patients was similar, although somewhat lower, than that previously reported by Frenneaux et al. (28)(i.e., 22% vs. 33%). These differences may be explained by the fact that the latter study used treadmill exercise testing (rather than cycloergometer) and probably comprised higher risk patients. Compared with those patients with a normal blood pressure response to exercise, patients with ABPR were also younger, more often experienced chest pain, had smaller left ventricular cavity size, and reached a lower degree of cardiovascular stress, as judged by the maximum double product achieved.

Hypertrophic cardiomyopathy–related mortality and ABPR

In our study, ABPR was associated with HCM-related death in patients ≤50 years, at the time of the exercise test, but was not a predictor of mortality in those patients >50 years. Among our study patients, only one of the nine HCM-related deaths occurred in patients >50 years of age, consistent with reports that HCM diagnosed later in life is usually associated with a benign prognosis (6,7). Thus, the survival analysis in the present study specifically describes that subset of HCM patients for whom risk stratification is most appropriate, that is, those ≤50 years. In the follow-up period subsequent to the exercise test, patients aged ≤50 years with ABPR showed a fourfold increase in premature HCM-related cardiovascular mortality (either suddenly or due to heart failure), compared with those with a normal blood pressure rise. However, the positive predictive accuracy of ABPR for mortality was low (i.e., 14%). This lack of predictability appears in large measure to be a consequence of the relatively uncommon occurrence of HCM-related death in our cohort, as reflected by the wide confidence interval evident in the multivariate survival analysis. Moreover, the uncertain significance of ABPR as a marker for the subsequent occurrence of clinical events is underlined by the fact that of the three patients in our study population who ultimately died suddenly, none collapsed during or just after exercise. Consequently, although we established a relationship between ABPR and increased cardiovascular mortality in patients ≤50 years of age, the positive predictive value of this single test is itself too low to allow identification of the high risk patient or modify clinical decision making. These observations substantiate the prior difficulties, shared by many authors, in defining subgroups of high risk HCM patients (2,6–10,20,22–25).

Identification of low risk patients

Conversely, ABPR showed a negative predictive accuracy of 95% for HCM-related mortality. This finding has important clinical implications because a normal blood pressure response during exercise testing appears to reliably identify those patients at lower risk. Therefore, as with other tests that have been proposed for risk stratification in patients with HCM (such as Holter ambulatory ECG) (23,25), assessment of the blood pressure response to exercise appears to be most useful as a screening study and an aid in identifying low risk subsets of HCM patients; indeed, such individuals are probably most patients within the overall HCM population (6). In this regard, a normal blood pressure response to exercise associated with the absenceof certain acknowledged risk factors—previous cardiac arrest or sustained ventricular tachycardia, family history of multiple sudden or other HCM-related deaths and malignant genotype, multiple-repetitive or prolonged bursts of nonsustained ventricular tachycardia on ambulatory Holter ECG, recurrent syncope and probably massive left ventricular hypertrophy (wall thickness ≥35 mm)—would appear to define those patients at low risk for an adverse clinical event, as suggested by Spirito et al. (6). Such patients probably do not require further aggressive diagnostic or therapeutic interventions and benefit most from reassurance regarding their prognosis (6,7).

Mechanisms and study limitations

The different prognostic implications of ABPR in HCM patients >50 years and ≤50 years may be the consequence of diverse underlying mechanisms in the two age groups. In younger patients, ABPR may reflect hemodynamic instability and thus suggest risk for a sudden cardiac event; however, in older patients, ABPR may simply reflect age-related cardiovascular changes which have little prognostic relevance. The mechanisms that may explain the abnormal pressure response to exercise in younger patients with HCM include: 1) impaired left ventricular filling secondary to tachycardia, associated with an inability to maintain an adequate stroke volume; or 2) an exaggerated fall in peripheral vascular resistance related to an abnormal ventricular mechanoceptor reflex, resulting in an inhibition of sympathetic tone in resistance vessels (28).

Exercise testing conducted in this large study population was used primarily as part of the clinical evaluation in an outpatient setting. Therefore, invasive hemodynamic measurements were impractical for our study patients, prohibiting a direct assessment of peripheral vascular resistance, cardiac output and consequently the precise physiologic mechanisms responsible for ABPR. A second limitation results from the small number of sudden cardiac deaths that occurred during the follow-up period. This low event rate probably reflects the relatively benign prognosis of HCM in our community-based cohort, as well as the administration of amiodarone to a sizeable subset of our patients. As a consequence, although ABPR proved to be an independent risk factor for combined sudden and heart failure-related HCM deaths in patients ≤50 years of age, we could not assess the prognostic significance of ABPR with regard to sudden and unexpected death alone.