Study Subjects

The diagnosis of HCM was confirmed in all patients by demonstrating a hypertrophied (≥15 mm wall thickness) non-dilated left ventricle in the absence of a secondary cause of hypertrophy. Patients were excluded if they had severe mitral regurgitation deemed not due to HCM, aortic stenosis, or more than mild aortic regurgitation.

All patients were assigned a New York Heart Association (NYHA) classification and completed the Minnesota Living with Heart Failure questionnaire (quality of life). All patients signed consent approved by the institutional review board of St. Luke’s-Roosevelt Hospital Center, New York, New York, for use of their clinical data for research purposes.

Echocardiography

The patients were referred outpatients studied while clinically stable in the course of treatment evaluation at the St. Luke’s-Roosevelt HCM Program. Echocardiograms with standard imaging planes were performed on commercially available echocardiographs, Acuson Sequoia C256 (Mountainview, CA) or Philips 5500 (Andover, MA). An experienced sonographer performed the two-dimensional echocardiograms using conventional broadband transducers with harmonic imaging. Continuous-wave Doppler was used to measure LVOT gradient from the apical four-chamber and apical three-chamber views to record maximum velocity parallel to the systolic flow of the LVOT. Care was taken to separate LVOT signal from that of mitral regurgitation by angling the beam medially so that it excluded the left atrium. Jets that began during isovolumetric systole, and were early peaking, and not concave-to-left, were assumed to be mitral regurgitation. The simplified Bernoulli equation was used to calculate LVOT gradient, ignoring the velocity proximal to the LVOT narrowing, as described previously, and the highest instantaneous gradient was reported. In all patients, LVOT gradient was measured in the supine position and during three separate Valsalva maneuvers. The patient was then asked to stand, and after 2 minutes of equilibration Doppler gradient was acquired over the next minute. To facilitate imaging while standing, the sonographer positioned the patient’s left arm on the patient’s head. This opened rib spaces and improved imaging.

Patients capable of exercising underwent treadmill testing with the Bruce protocol and had gradients acquired after exercise. Patients were excluded from exercise if they had orthopedic disabilities or such severe heart failure symptoms that they were deemed incapable of performing even a modified, reduced treadmill protocol, or for rest gradients ≥ 80 mm Hg. Beta-blockers were held the morning of stress test, whereas other medications were continued. After stress exercise, continuous-wave Doppler gradients from apical views were acquired in supine left lateral decubitus position as soon as possible after completion of exercise within 30 to 60 seconds and then again 3 minutes later.

Measurements

Maximum LV wall thickness and LV end-diastolic dimensions were measured from parasternal long- and short-axis views from the two-dimensional echocardiogram as previously described. Two experienced readers selected and measured the highest gradient for each provocation. The mean of the two observers’ data was the reported gradient. When there was no mitral-septal contact and LVOT velocities were ≤ 1.8 m/sec, a zero gradient was assigned. Interobserver variability was assessed by comparing the standing gradients measured by the 2 readers in 36 patients. Intraobserver variability also was assessed. Mitral regurgitation was scored from 0 to 3 by the jet area/left atrial area method.

Statistics

SPSS 10.0.5 (SPSS Inc., Chicago, IL) was used for statistical analysis. Data are presented as mean value ± standard deviation. Skewed distributions are presented as median (range). To compare the Doppler gradients and the increase in gradients with rest, after standing, Valsalva, and post-exercise, analysis of variance with the Kruskal–Wallis test was used. Pairwise comparisons were performed using the Wilcoxon rank-sum test with Bonferroni corrections for multiple analyses. Correlations were performed using Pearson’s correlation coefficient. A P value < .05 was considered significant. For interobserver and intraobserver variability, Cronbach’s alpha analysis was used.

Change in Gradient after Standing

Resting gradients were 36 ± 39 mm Hg, median 25 mm Hg (range 0–205 mm Hg), skew 2.1 ± 0.24. Fifty-six (57%) of the 98 patients had resting gradients < 30 mm Hg and would thus be characterized as nonobstructive at rest.

Compared with the supine position, standing provoked an increase in LVOT gradient ≥ 10 mm in 51 of 98 patients (52%). Median gradient increased from 25 to 44 mm Hg (range 0–309), an increase of 76% ( P < .001). The distribution of the absolute increase in gradient with standing is shown in Figure 1 . In 47 of 98 patients (48%), gradient was essentially unchanged after standing, increasing < 10 mm Hg; in 18 of 98 patients (18%) gradients increased by 10 to 30 mm Hg, and in 33 of 98 patients (34%), standing provoked a gradient increase of > 30 mm Hg. Of the 56 patients who had supine resting gradients < 30 mm Hg, 23 of 56 (41%) had an increase to ≥ 30 mm Hg with standing, thus placing them in the domain of obstructive HCM. Of the 75 patients who had supine resting gradients < 50 mm Hg, 34 of 74 (46%) had an increase to ≥ 50 mm Hg.

Increase in Doppler echocardiographic gradients after standing and Valsalva maneuver in 98 patients with HCM. For standing gradient, the distribution of increase is skewed to the left with fewer patients showing a large increase in gradient (positive skew 1.73, with 2 × SES = 0.48).

Valsalva and Exercise Gradients

Valsalva gradients were a median of 64 mm Hg (range 0–256 mm Hg), again higher than supine ( P < .001) and standing ( P = .004). There was a similar positive skew in the distribution of gradient increase. Figure 1 shows the distribution of the increase in gradient after standing and after Valsalva. Valsalva provoked a greater increase in gradient than standing, a median increase of 36 versus 12 mm Hg ( P = .002). However, in 29 of 98 patients (30%), the standing gradient was higher than the Valsalva Doppler gradient. Figure 2 shows the Doppler tracings in one such patient.

Patient with obstructive HCM and latent obstruction. LV outflow gradient supine is 19 mm Hg. Gradient increases to 49 mm Hg after Valsalva, to 74 mm Hg after standing, and to 144 mm Hg after treadmill exercise. In 30% of cases, we found a higher standing than Valsalva gradient, as in this patient. Of the 56 patients with resting gradients < 30 mm Hg (57%), standing provoked a gradient ≥ 30 mm Hg in 23 (41%), thus placing them in the domain of obstructive HCM.

Among the 53 patients who had exercise provocation, gradients increased to a median of 100 mm Hg (range 0–256 mm Hg), a median increase of 65 mm Hg after exercise compared with rest ( P < .001) and a greater increase than after Valsalva ( P < .001). Median gradients were 25 mm Hg at rest (range 0–205 mm Hg), 44 mm Hg after standing (range 0–309 mm Hg), 64 mm Hg after Valsalva (range 0–256 mm Hg) in the 98 patients, and 100 mm Hg in the 53 patients after exercise (range 0–256 mm Hg) ( P < .001). Overall, after the provocations, 89 of 98 patients (91%) in this consecutive referred cohort had obstruction ≥ 30 mm Hg demonstrated either at rest, after standing, after Valsalva, or after exercise. Gradients ≥ 50 mm Hg were present in patients taking negatively inotropic medications in approximately 30% of individuals when measured supine, 50% when measured standing, and 70% when measured after Valsalva and exercise, in part because of our laboratory practice of holding beta-blockers the morning of stress tests. Nineteen patients (19%) had persistent disabling symptoms and gradients despite up-titration of medication, and underwent surgical septal myectomy at St. Luke’s-Roosevelt 12 ± 18 months later.

Correlations

Workload achieved in Mets achieved correlated negatively with quality of life scores ( r = −0.44, P = .002) and NYHA class ( r = −0.46, P = .001). Exercise time also correlated negatively with quality of life ( r = −0.45, P = .002) and NYHA class ( r = −0.44, P = .002). There was no correlation of quality of life scores or NYHA class with rest, standing, Valsalva, or exercise gradients. For each individual patient, there was correlation between the increase in gradient from supine to standing and the increase in gradient after Valsalva ( r = 0.40 P < .001). There was also a correlation between the increase in gradient from supine to standing and the increase after exercise ( r = 0.28, P < .05). For measurement of standing Doppler gradient interobserver variability was low (Cronbach’s alpha = 0.97). The average difference between the 2 observers was 3 mm Hg. Intraobserver variability similarly was low (alpha = 0.98). One patient experienced ventricular tachycardia on the treadmill that was terminated by a low-level discharge from his implanted defibrillator.