Abstract
Aims
We sought to evaluate associations between baseline sphericity index (SI) and clinical outcome, and changes in SI after coronary artery bypass graft surgery (CABG) with or without surgical ventricular reconstruction (SVR) in ischemic cardiomyopathy patients enrolled in the SVR study (Hypothesis 2) of the Surgical Treatment for Ischemic Heart Failure (STICH) trial.
Methods and results
Among 1,000 patients in the STICH SVR study, we evaluated 546 patients (255 randomized to CABG alone and 291 to CABG+SVR) whose baseline SI values were available. SI was not significantly different between treatment groups at baseline. After 4 months, SI had increased in the CABG+SVR group, but was unchanged in the CABG alone group (0.69 ± 0.10 to 0.77 ± 0.12 versus 0.67 ± 0.07 to 0.66 ± 0.09, respectively; P < 0.001). SI did not significantly change from 4 months to 2 years in either group. Although LV end-systolic volume and ejection fraction improved significantly more in the CABG+SVR group compared to CABG alone, the severity of mitral regurgitation significantly improved only in the CABG alone group and estimated LV filling pressure (E/A ratio) increased only in the CABG+SVR group. Higher baseline SI was associated with worse survival after surgery (hazard ratio = 1.21, 95% confidence interval = 1.02−1.43; P = 0.026). Survival was not significantly different by treatment strategy.
Conclusion
Although SVR was designed to improve LV geometry, SI worsened after SVR despite improved LV ejection fraction and smaller LV volume. Survival was significantly better in patients with lower SI regardless of treatment strategy. (THE STICH TRIAL: Surgical Treatment for Ischemic Heart Failure trial; NCT00023595)
Keywords: Sphericity index, Surgical ventricular reconstruction and STICH
Introduction
In patients with ischemic cardiomyopathy, left ventricular (LV) remodeling occurs preferentially along the short axis of the LV and results in a dilated ventricle with a more spherical shape, which is associated with poor outcomes.1,2 Surgical ventricular reconstruction (SVR) is a technique to attempt to reverse LV remodeling and to restore the more efficient geometry of LV in patients with previous large anterior myocardial infarction and apical akinesia or dyskinesia.3 However, the data from the Surgical Treatment for Ischemic Heart Failure (STICH) trial did not demonstrate survival benefit from adding SVR to coronary artery bypass grafting (CABG) compared to CABG alone in that patient population, despite improved LV ejection fraction (LVEF) and smaller LV volumes with SVR.4 Moreover, a subgroup analysis based on baseline echocardiography measurements suggests that SVR may improve clinical outcome in patients with an early stage of LV remodeling and LV end-systolic volume (LVESV) index of ≤60 ml/m2, but not in patients with a larger LV.5,6
Sphericity index (SI), which is a ratio of LV short-axis to long-axis dimension, has been used to evaluate geometry of LV and the higher SI value indicates more globular shape of LV and poor prognosis.2 Therefore, the purpose of this report is to determine how SVR affects the sphericity index (SI) in patients with ischemic cardiomyopathy enrolled in the SVR study (Hypothesis 2) of the STICH trial and what impact the baseline SI had on the clinical outcome of the patients after CABG with or without SVR.
Methods
Patient selection
Between September 2002 and January 2006, patients with an LVEF ≤35%, apical dyssynergy, and coronary artery disease amenable to CABG were enrolled in the SVR hypothesis of the STICH trial and randomized to CABG+SVR or CABG alone. Patients were recruited from 122 clinical sites in 26 countries. The qualifying LVEF for enrollment was determined by the clinical sites using any available imaging modalities within 3 months of enrollment. More detailed inclusion and exclusion criteria and randomization strata for the STICH trial have been published elsewhere.4,7 The median duration of follow-up of the patients was 4 years. The primary end-points in the current analysis were over-all mortality and change of SI at 4-month follow-up. In this analysis, we excluded patients if their baseline SI was unavailable. To insure that the subset of patients for the analysis of changes in SI and other echocardiographic variables was not compromised by early surgical death (early post-surgical death is more common in patients with higher risks such as larger LV size and higher SI at baseline), we defined a subgroup of patients who underwent surgery and remained alive to undergo follow-up imaging of the LV size and shape using paired evaluations of SI in patients with adequate quality of echocardiography allowing LV volume measurement.6
Echocardiography
Transthoracic echocardiography (TTE) was performed at baseline, 4 months (between 2 and 6 months) and 2 years (between 18 and 30 months) following surgery and TTE images were submitted to the Echocardiography Core Laboratory for analysis. All TTE measurements were performed by the Echocardiography Core Laboratory according to standardized methods as recommended by the American Society of Echocardiography.8,9 All measurements and analyses were performed without knowledge of treatment assignment, clinical or other laboratory data. Echocardiographic measurements were acquired and averaged over 3 cardiac cycles if sinus rhythm was present, and 3 to 5 cardiac cycles for patients in atrial fibrillation.
Left ventricular dimensions and sphericity index
The LV SI, which was a predefined echocardiographic variable in STICH trial, was calculated as the ratio of the LV short-axis dimension (numerator) and the long-axis dimension (denominator) measured at end-diastolic period. Therefore, the more spheric LV becomes, the higher the SI. LV minor short-axis dimension was measured from the 2D parasternal long-axis view of the LV near the junction of the head of the papillary muscle and chordae. Long-axis dimension of the LV was measured from the apical 4-chamber view from the mitral annulus to the apex (Figure 1). For follow-up SI after SVR, long-axis dimension was measured from mid-portion of mitral annulus to the center of LV apex in the apical 4-chamber view, which might be different from longest dimension of LV.
Figure 1.
Measurement of sphericity index (SI). LV short-axis dimension (SaD) was measured near the junction between the papillary muscle head and the chordae in parasternal long-axis view and long-axis dimension (LaD) from the mitral annulus to the apex by drawing a line at the mid-portion of annular plane to the center of LV apex in the apical 4-chamber view. SI was calculated at baseline (left panels) and 4 months after surgery (right panels) measuring left ventricular (LV) parasternal long-axis view (upper panels) and apical 4-chamber view (lower panels) at end-diastole. Note that despite slight decrease in SaD, greater reduction in LaD results in increase of SI at 4 months.
Left ventricular volume and ejection fraction measurement
LVEF was measured by the Simpson’s biplane volumetric method whenever possible. Either a combination of apical 4- and 2-chamber views (preferentially) or a combination of apical 4-chamber and long-axis views was used. If 2 apical views were not available, only 1 apical view was used for the Simpson’s single plane method. The LV endocardial border was traced contiguously from 1 side of the mitral annulus to the other, excluding the papillary muscles and trabeculations. LVEF was determined from LV volumes accordingly, which were indexed by body surface area.
Statistical analyses
Descriptive statistics on baseline data were reported as mean±standard deviation, median (interquartile ranges) or number (percentage), as appropriate. Two-group comparisons of baseline characteristics between the two types of surgeries were tested using the Student’s t-test or Wilcoxon Rank Sum test for continuous data or a Chi-Square test for categorical variables. All comparisons of treatment strategy were analyzed based on the intent-to-treat approach. As the primary variable of interest, a measured value of SI at baseline was required for inclusion in these analyses. Since a possible selection bias could limit the generalizability of results from this subset, analyses comparing the characteristics of those with and without an SI value were performed (see Supporting Information, Table S1). For all analyses, P < 0.05 was considered statistically significant.
For SI and other echo parameters, the change in value between baseline and 4-month follow-up was assessed on the subset in which both measures were available. Likewise, the change between 4-month and 2-year follow-up values was analyzed, though fewer subjects with those paired measures were available. We tested for a significant change between paired measures using a paired t-test, and for a significant difference in the change between surgery groups using a two-sample t-test. Pearson correlation coefficients were used to assess whether baseline LVESV index was associated with baseline SI or with changes in SI from baseline to 4 months within surgery groups and whether changes in echocardiographic variables over time were linearly related to changes in SI. For these analyses, grade of mitral regurgitation was quantified as a numerical variable as follows: 0 for none, 1 for mild, 2 for moderate, 3 for moderately severe and 4 for severe mitral regurgitation. Likewise, grade of diastolic dysfunction was coded as a numerical variable with a grade of 0 to 4. An indeterminate grade of diastolic function or mitral regurgitation was treated as missing data, thereby excluding those subjects from the corresponding analyses.
The influence of SI on overall survival was evaluated with Cox proportional hazards regression, both unadjusted and adjusted for conventional confounders including age, sex, and body mass index. To study whether the effect of SI on the endpoint was differential with respect to surgery type, we fit a model with an interaction term between SI and surgery type and tested its significance. A linear effect of SI was assumed in these regression models since a p-spline plot revealed no clear evidence of a non-linear relationship between the log of the mortality hazard function and numerical values of SI.10 For descriptive purposes, survival was estimated by the Kaplan-Meier method for categories of SI, divided at the median, and plotted over time.
Results
Baseline echocardiographic data
A total of 1000 patients were enrolled in the SVR arm, and randomized to CABG alone (n = 499) or CABG+SVR group (n = 501). Of these, 937 underwent baseline TTE, 724 at 4 months and 561 at 2 years following surgery (Figure 2). Baseline characteristics for the study population have been previously published.4 After exclusion of 391 subjects whose image was poor for measurement of SI (unable to measure LV long-axis [n = 222] or short-axis dimension [n = 265]; 96 subjects were missing both), we included 546 study subjects whose baseline SI was available. Study subjects were younger, less frequently male, had lower body mass index and were more frequently randomized to CABG+SVR group compared to those without a baseline SI measure (see Supporting Information, Table S1). Also, despite no significant differences in baseline LVEF, those patients with baseline SI had a slightly larger LV volume, increased E/A ratio, shorter deceleration time, and more significant mitral regurgitation and diastolic dysfunction compared to patients in which baseline SI could not be measured.
Figure 2.
Flow diagram for analysis of the current study. *Study subjects included for outcome analysis for sphericity index. †Study subjects included for the changes in sphericity index from baseline to 4-month follow-up. ‡Study subjects included for the changes in sphericity index from 4-month to 2-year follow-up. STICH, Surgical Treatment for Ischemic Heart Failure; CABG, coronary artery bypass graft surgery; SVR, surgical ventricular reconstruction; TTE, transthoracic echocardiography; SI, sphericity index; LV, left ventricular; LAD, long-axis dimension; SAD, short-axis dimension; FU, follow-up.
For the 546 subjects included in the primary analyses (n = 255 with CABG alone, n = 291 with CABG+SVR), baseline echocardiographic data according to the treatment group are shown in Table 1. At baseline, there were no significant differences in LVEF, LV volumes, SI, or echocardiographic parameters for diastolic function between the CABG alone group compared with the CABG+SVR group.
Table 1.
Baseline characteristics of study subjects in the STICH trial SVR evaluation with available baseline SI according to treatment strategy
| Variable | Overall | CABG | CABG+SVR | P Value | |||
|---|---|---|---|---|---|---|---|
|
| |||||||
| N | Mean±SD or n (%)* | N | Mean±SD or n (%)* | N | Mean±SD or n (%)* | ||
| Age at randomization (years) | 546 | 60.6±9.6 | 255 | 60.4±9.6 | 291 | 60.7±9.6 | 0.78 |
| Male gender | 546 | 450 (82.4%) | 255 | 204 (80.0%) | 291 | 246 (84.5%) | 0.16 |
| Body mass index (kg/m2) | 546 | 26.9±4.1 | 255 | 27.2±3.9 | 291 | 26.7±4.2 | 0.14 |
| Body surface area (m2) | 546 | 1.92±0.21 | 255 | 1.93±0.20 | 291 | 1.92±0.21 | 0.33 |
| Creatinine (mg/dL), Median (Q1, Q3) | 545 | 1.06 (0.90, 1.26) | 254 | 1.05 (0.90, 1.21) | 291 | 1.06 (0.91, 1.28) | 0.58 |
| LVEF (%) | 509 | 29.4±8.3 | 238 | 29.3±8.1 | 271 | 29.4±8.5 | 0.83 |
| LVEDV (ml) | 509 | 228.0±70.6 | 238 | 227.8±67.1 | 271 | 228.2±73.6 | 0.94 |
| LVESV (ml) | 509 | 163.5±61.7 | 238 | 163.6±59.9 | 271 | 163.5±63.4 | 0.98 |
| LVEDV index (ml/m2) | 509 | 119.3±36.3 | 238 | 118.4±33.8 | 271 | 120.1±38.3 | 0.60 |
| LVESV index (ml/m2) | 509 | 85.6±32.1 | 238 | 85.0±30.4 | 271 | 86.2±33.5 | 0.69 |
| LVEDD or short-axis dimension(cm) | 546 | 6.37±0.82 | 255 | 6.37±0.79 | 291 | 6.36±0.85 | 0.90 |
| LVESD (cm) | 515 | 5.35±0.95 | 240 | 5.35±0.93 | 275 | 5.35±0.96 | 0.94 |
| LV long-axis dimension (cm) | 546 | 9.35±0.97 | 255 | 9.38±0.93 | 291 | 9.32±1.00 | 0.45 |
| Sphericity index | 546 | 0.68±0.09 | 255 | 0.68±0.08 | 291 | 0.69±0.10 | 0.42 |
| RWT | 527 | 0.30±0.08 | 252 | 0.29±0.08 | 275 | 0.30±0.08 | 0.47 |
| Left atrial volume (ml) | 378 | 81.6±29.0 | 175 | 81.6±30.0 | 203 | 81.6±28.2 | 0.99 |
| E velocity (m/sec) | 479 | 0.76±0.24 | 222 | 0.76±0.25 | 257 | 0.75±0.24 | 0.63 |
| A velocity (m/sec) | 455 | 0.66±0.25 | 212 | 0.66±0.25 | 243 | 0.66±0.26 | 0.89 |
| E/A ratio, median (Q1, Q3) | 454 | 1.13 (0.71, 1.80) | 212 | 1.16 (0.72, 1.78) | 242 | 1.03 (0.71, 1.83) | 0.62 |
| Deceleration time (msec) | 445 | 177.5±49.4 | 208 | 181.4±51.0 | 237 | 174.2±47.8 | 0.12 |
| e′, septal (m/sec) | 294 | 0.05±0.02 | 138 | 0.05±0.02 | 156 | 0.05±0.02 | 0.73 |
| e′, lateral (m/sec) | 278 | 0.06±0.03 | 133 | 0.06±0.03 | 145 | 0.06±0.03 | 0.12 |
| E/e′ septal, Median (Q1, Q3) | 271 | 15.6 (11.4, 20.0) | 128 | 15.9 (12.0, 21.8) | 143 | 15.0 (10.0, 20.0) | 0.40 |
| E/e′ lateral, Median (Q1, Q3) | 263 | 12.5 (9.3, 16.7) | 124 | 12.9 (10.0, 17.8) | 139 | 12.0 (8.6, 16.7) | 0.16 |
| MR grade | 539 | 252 | 287 | 0.12 | |||
| None | 114 (21.2%) | 63 (25.0%) | 51 (17.8%) | ||||
| Mild | 261 (48.4%) | 118 (46.8%) | 143 (49.8%) | ||||
| Moderate | 92 (17.1%) | 42 (16.7%) | 50 (17.4%) | ||||
| Moderate to severe | 36 (6.7%) | 17 (6.7%) | 19 (6.6%) | ||||
| Severe | 17 (3.2%) | 8 (3.2%) | 9 (3.1%) | ||||
| Indeterminate | 19 (3.5%) | 4 (1.6%) | 15 (5.2%) | ||||
| Diastolic function grade | 545 | 254 | 291 | 0.87 | |||
| Normal | 2 (0.4%) | 1 (0.4%) | 1 (0.3%) | ||||
| 1 | 142 (26.1%) | 67 (26.4%) | 75 (25.8%) | ||||
| 2 | 191 (35.0%) | 91 (35.8%) | 100 (34.4%) | ||||
| 3 | 138 (25.3%) | 60 (23.6%) | 78 (26.8%) | ||||
| 4 | 1 (0.2%) | 1 (0.4%) | 0 (0.0%) | ||||
| Indeterminate | 71 (13.0%) | 34 (13.4%) | 37 (12.7%) | ||||
| PASP (mmHg) | 151 | 42.8±15.2 | 69 | 40.6±14.3 | 82 | 44.7±15.7 | 0.10 |
| MV repair/replacement | 546 | 104 (19.0%) | 255 | 42 (16.5%) | 291 | 62 (21.3%) | 0.15 |
Mean±standard deviation or count (percentage) shown unless otherwise noted. A, late mitral inflow velocity; CABG, coronary artery bypass graft surgery; E, early mitral inflow velocity; e′, early mitral annular velocity; LV, left ventricle; LVEDD, LV end-diastolic dimension; LVEDV, LV end-diastolic volume; LVEF, LV ejection fraction; LVESD, LV end-systolic dimension; LVESV, LV end-systolic volume; MR, mitral regurgitation; MV, mitral valve; PASP, pulmonary artery systolic pressure; RWT, relative wall thickness; SVR, surgical ventricular reconstruction.
Changes in LV volume, ejection fraction and sphericity index
Changes in echocardiographic variables from baseline to 4-month follow-up are presented according to the treatment group in Table 2, Supporting Information Table S2, and Figure 3. Baseline characteristics of the patients who were excluded from this analysis due to not having a 4-month follow-up SI measured are also shown in the Supporting Information, Table S3. There were no significant differences in mean time from baseline to 4 months echocardiography (4.3 ± 0.5 months for CABG alone and 4.4±0.6 months for CABG+SVR, P = 0.06), and from baseline to 2 years follow-up echocardiography (24.6 ± 1.2 months for CABG alone and 24.6 ± 1.1 months for CABG+SVR, P = 0.87) between groups. LVESV index was significantly reduced at 4 months after surgery in both the CABG+SVR (83.2 ± 33.5 ml/m2 to 67.7 ± 27.8 ml/m2, P < 0.001) and CABG alone (84.2 ± 28.0 ml/m2 to 79.2 ± 32.2 ml/m2, P = 0.037) groups, but the change was significantly greater in the CABG+SVR group (P = 0.002). LVEF increased significantly in the CABG+SVR group (29 ± 9% to 35 ± 11%, P < 0.001) but not in the CABG alone group (30 ± 8% to 32 ± 10%, P = 0.11). Likewise, LV long-axis dimension was significantly decreased in the CABG+SVR group (9.3 ± 1.1 cm to 8.2 ± 1.0 cm, P < 0.001) but not in the CABG alone group (9.5 ± 0.9 cm to 9.4 ± 0.9 cm, P = 0.10). In contrast, LV short-axis dimension (LV end-diastolic dimension) was significantly decreased in the CABG alone group (6.4 ± 0.7 cm to 6.2 ± 0.8 cm, P = 0.007) but not in the CABG+SVR group (6.3 ± 0.8 cm to 6.2 ± 0.8 cm, P = 0.08), although these changes was not significantly different between groups (P = 0.37). Accordingly, SI worsened after 4 months in the CABG+SVR group (0.69 ± 0.10 to 0.77 ± 0.12, P < 0.001) but did not significantly change in the CABG alone group (0.67 ± 0.07 to 0.66 ± 0.09, P = 0.24). In general, among the subsets who additionally had 2-year data, echocardiographic measurements did not change significantly from 4 months to 2 years, with the exception of a slight increase in LV end-diastolic dimension in the CABG+SVR group (6.07 ± 0.70 cm to 6.22 ± 0.72 cm, P = 0.011, Figure 3).
Table 2.
Changes in echocardiographic variables from baseline to 4 months according to treatment strategy
| Variable | CABG | CABG+SVR | P Value§ | ||
|---|---|---|---|---|---|
|
| |||||
| N | Mean±SD* | N | Mean±SD* | ||
| Sphericity index | 111 | −0.01±0.08 | 117 | 0.08±0.10‡ | <0.001 |
| LV long-axis dimension (cm) | 111 | −0.12±0.77 | 117 | −1.1±1.1‡ | <0.001 |
| LVEDD or short-axis dimension (cm) | 111 | −0.17±0.67‡ | 117 | −0.10±0.61 | 0.37 |
| LVESD (cm) | 98 | −0.17±0.75† | 108 | −0.16±0.80† | 0.96 |
| LVEF (%) | 101 | 1.8±11.2 | 94 | 5.6±9.6‡ | 0.012 |
| LVEDV (ml) | 101 | −9.4±45.3† | 94 | −26.2±52.0‡ | 0.017 |
| LVESV (ml) | 101 | −9.1±45.9† | 94 | −29.1±44.0‡ | 0.002 |
| LVEDV index (ml/m2) | 101 | −5.1±24.3† | 94 | −14.0±27.3‡ | 0.017 |
| LVESV index (ml/m2) | 101 | −5.1±24.1† | 94 | −15.5±23.0‡ | 0.002 |
| RWT | 107 | 0.03±0.10‡ | 110 | 0.01±0.10 | 0.09 |
| Left atrial volume (ml) | 57 | −1.5±23.6 | 44 | 1.0±30.6 | 0.63 |
| E velocity (m/sec) | 88 | 0.10±0.30‡ | 96 | 0.17±0.34‡ | 0.16 |
| A velocity (m/sec) | 80 | 0.11±0.33‡ | 88 | −0.03±0.26 | 0.001 |
| E/A ratio | 80 | −0.09±0.89 | 88 | 0.45±1.2‡ | 0.001 |
| Deceleration time (msec) | 80 | 16.2±63.3† | 83 | −4.7±74.0 | 0.06 |
| MR grade | 102 | −0.34±1.1‡ | 106 | −0.07±1.00 | 0.06 |
| Diastolic function grade | 78 | −0.01±0.80 | 90 | 0.18±0.92 | 0.16 |
P value of change within group is from a paired t-test:
indicates P < 0.05
indicates P < 0.01.
P value of comparison of changes of variables between treatment strategies from two-sample t-test. For abbreviations, see previous table.
Figure 3.
Changes in echocardiographic variables. Mean values along with 95% confidence intervals of echocardiographic variables at baseline and 4 month follow-up (left side in each graph) and at 4-month and 2-year follow-up (right side in each graph) are noted according to the treatment groups. Upper panels show the changes in sphericity index (SI) together with those of left ventricular (LV) short-axis and long-axis dimension, which are the numerator and denominator of SI, respectively. Lower panels show changes in LV end-systolic volume index (LVESV index), LV ejection fraction (LVEF) and deceleration time of early mitral inflow velocity. Note that even though reduced LVESV index and improved LVEF was noted, SI got worse in CABG+SVR group. Deceleration time was increased only in CABG group. *indicate P < 0.05 from a paired t-test within treatment group. Blue lines indicate CABG alone group and red lines CABG+SVR group. CABG, coronary artery bypass graft surgery; SVR, surgical ventricular reconstruction.
In both treatment arms, baseline levels of SI and indexed LVESV were positively correlated (Figure 4). However, the change in SI was not significantly associated with baseline LVESV index.
Figure 4.
Associations of baseline LV end-systolic volume (LVESV) index with baseline sphericity index (SI) (upper panels) and change in SI from baseline to 4 months within treatment group (lower panels). Pearson correlation coefficients along with p values are shown at the top of each panel. Linear regression lines are also shown for each panel.
Changes in mitral regurgitation and diastolic function
Mitral valve repair and/or replacement was performed during the surgery in 62 (21.3%) patients of the CABG+SVR group, which was not significantly different compared with CABG alone group (42 [16.5%] patients, P = 0.15). However, mitral regurgitation grade was significantly decreased in the CABG group (1.3 ± 1.1 to 0.9 ± 0.7, P = 0.003) but did not significantly change in the CABG+SVR group (1.2 ± 0.9 to 1.1 ± 1.0, P = 0.50). Despite the within group change of mitral regurgitation grade, the difference in change in mitral regurgitation between treatment groups was only of marginal statistical difference (P = 0.06).
There were significant differences in changes in diastolic function parameters from baseline to 4 months according to the treatment strategy. In particular, E/A ratio which estimates LV filling pressure increased significantly in the CABG+SVR group (1.3 ± 0.9 to 1.8 ± 1.1, P < 0.001), suggesting a higher diastolic filling pressure, but did not significantly change in the CABG alone group (1.4 ± 0.9 to 1.3 ± 0.8, P = 0.35). These trends did not change when we excluded the patients who underwent mitral valve repair and/or replacement from the paired comparisons. Interestingly, changes in the grade of mitral regurgitation and LV diastolic dysfunction had weak but statistically significant correlations with change in SI from baseline to 4 months (r = 0.17, P = 0.025 and r = 0.16, P = 0.020, respectively) among all subjects combined.
Sphericity index and outcome
Lower baseline SI was associated with better overall survival after surgery (hazard ratio = 1.21, 95% confidence interval = 1.02−1.43, P = 0.026, Figure 5, Table 3). Survival was not affected by treatment strategy (P = 0.94) nor was the effect of SI on survival different by treatment group (P = 0.27). Neither SI at 4 month follow-up or change in SI from baseline to 4 month had a significant association with overall survival after 4 month in landmark survival analyses.
Figure 5.
Kaplan-Meier curves for overall survival according to the median baseline sphericity index (SI) in overall study subjects (left panel) and within treatment strategy (right panel). Age, sex, and body mass index-adjusted hazard ratio from Cox-regression analysis is 1.21 (95% confidence interval = 1.02−1.43, P = 0.026) per 0.1 increase of SI. However, there was no significant interaction between treatment and SI in predicting overall survival (P = 0.27).
Table 3.
Cox proportional hazard models for overall survival
| Variable | N of patients | N of events | Unadjusted
|
Age, sex and BMI-Adjusted
|
||
|---|---|---|---|---|---|---|
| HR (95% CI) | P Value | HR (95% CI) | P Value | |||
| Sphericity index, per 0.1 | 546 | 156 | 1.20 (1.01−1.42) | 0.041 | 1.21 (1.02−1.43) | 0.026 |
| Treatment CABG+SVR vs. CABG | 546 | 156 | 0.98 (0.72−1.35) | 0.92 | 0.99 (0.72−1.35) | 0.94 |
| Sphericity index at 4 months, per 0.1 | 228 | 39 | 1.13 (0.86−1.48) | 0.39 | 1.18 (0.85−1.63) | 0.33* |
| Change in sphericity index from baseline to 4-month echo, per 0.1 | 228 | 39 | 1.15 (0.85−1.57) | 0.37 | 1.13 (0.83−1.55) | 0.44 |
There was no significant interaction with treatment and sphericity index in predicting overall survival (P = 0.27).
Adjustment was made for baseline sphericity index as well. N, number; BMI, body mass index; HR, hazard ratio; CI, confidence interval; for other abbreviations, see previous tables.
Discussion
The major findings of the current data are that survival was worse in patients with increased ventricular sphericity regardless of treatment strategy, and that SI increased in the SVR patients most likely due to the significant shortening of LV long-axis without shortening of the short-axis of LV.
Previous studies have identified larger LV volumes, restrictive diastolic filling, advanced heart failure symptoms, and abnormal LV geometry as predictors of poor outcomes in patients undergoing SVR.11–15 STICH is the largest surgical trial in patients with ischemic cardiomyopathy directly comparing the impact of CABG alone versus CABG+SVR. The results of STICH demonstrate that increased SI was associated with worse survival, independent of age, sex, and body mass index for the patients with ischemic cardiomyopathy undergoing surgical revascularization. The results of STICH also suggest that the negative impact of abnormal LV geometry on mortality is independent of and not significantly different whether SVR is added or not to CABG.
Initially, SVR was believed to reverse LV remodeling by eliminating the akinetic or dyssynergic zone of LV and therefore reducing LV cavity size with a more efficient shape. However, there were controversies about the usefulness of this procedure with the only previous study of a small number of patients, showing that SI became worse after SVR.14,16 In this analysis of STICH data, LV end-systolic and diastolic volumes were reduced and LVEF improved after CABG+SVR as expected. The decrease of LV volume after SVR was likely due to reduction in LV long-axis dimension as there was only an insignificant decrease in short-axis dimension. Although LV volume was reduced after SVR, its shape most likely became more globular and SI increased at follow-up. Considering that diastolic function is usually more related to the longitudinal motion of the LV, findings that the reduction in LV long-axis dimension and increased SI were associated with worsening grades of diastolic dysfunction are noteworthy.17 Indeed, as the LV becomes more globular, LV wall tension generally increases and is followed by more severe diastolic dysfunction and/or mitral regurgitation, which was consistent in the current data.14,18,19 Increased LV wall tension could also have adverse effects on myocardial blood flow.20 This counter-intuitive change in LV shape towards a less efficient globular shape after SVR was evident irrespective of the baseline LVESV index (Figure 4). In contrast, SI was not significantly changed in the CABG alone group although there was small but statistically significant reduction in LV short-axis dimension.
Although a smaller LV volume and better LVEF usually indicate better outcome, after SVR these changes were not translated into a survival benefit in STICH. This lack of benefit in survival with the addition of SVR may have been a result of the LV becoming more spherical (with an increase in SI) and worsening of LV diastolic dysfunction or filling pressure (as assessed by E/A ratio) in the CABG+SVR group and a significant improvement in the severity of mitral regurgitation being documented only in the CABG alone group. Indeed, worse diastolic dysfunction and significant mitral regurgitation are well-known clinical predictors of poor outcomes in patients with heart failure, such that the beneficial effects in terms of LV volume and LVEF in patients with CABG+SVR may have been offset by the worse effects on diastolic function and mitral regurgitation.
Previously our report from STICH trial showed that patients with smaller LV had a benefit from SVR in the contrary to the previous expectation that SVR would be useful the patients with advanced LV remodeling.6 This subgroup of patients with early stage of LV remodeling could partly explain why there were no significant differences in clinical outcome between CABG+SVR and CABG alone groups in spite of worse geometry, diastolic function and mitral regurgitation. However, as there was no significant interaction between treatment strategies and baseline SI on overall survival, it is difficult to say whether outcome of CABG+SVR was even worse in the patients with higher baseline levels of SI compared to those with lower values. These insignificant results might be from small number of patients evaluated in 4-month echocardiography or relatively short follow-up periods. However, our data might support that artificial modification of LV geometry by SVR could not change the clinical outcome of the patients with ischemic cardiomyopathy.
Interestingly only few cases with higher baseline LVESV index (>90 ml/m2) showed reduced SI at 4 months after surgery and those who experienced reduction of SI had relatively small LVESV index at baseline (Figure 4). This finding might give an insight about why clinical course after SVR was better in the patients with small LVESV index or less advanced LV remodeling in the previous reports.6
Study limitations
There were several limitations in the current analysis. First, baseline SI measurement was available in only 55% of the patients who were originally assigned to the SVR arm of the STICH trial. There was also considerable loss of patients for the follow-up measurements of the echocardiographic variables. One reason for this is that LV dimensions were measured in the Echocardiography Core Laboratory only when the echocardiography images taken for that measurement were completely satisfactory. As the baseline characteristics of study subjects in which baseline SI measurements were available had slight but statistically significantly different baseline characteristics from those not included (see Supporting Information, Table S1), the study results may not be generalizable to the entire set of patients enrolled into the STICH trial. Nevertheless, since there were no noticeable group imbalances in baseline echocardiographic characteristics between treatment groups in patients for whom an SI was obtainable, this selection did not appear to bias the group comparability of our primary analyses. A number of additional cases were excluded from the paired comparison analyses of SI between baseline and 4 months and between 4 months and 2 years. This was mainly due to the frequency of very poor echocardiographic windows for good images after surgery and the fact that we could not evaluate the changes in those who died during the follow-up. This resulted in lack of sufficient statistical power to detect the significant associations between survival and SI at 4 months or change of SI from baseline to 4 months. Therefore, it is difficult to conclude whether this lack of significant association with survival was due to a small number of follow-up patients or no actual clinical impact of increased SI after SVR on overall survival in this study. Follow-up echocardiography was performed in relative wide ranges of time periods; between 2 and 6 months for the 4-month and between 18 and 30 months for the 2-year follow-up. This wide range of time point might have result in deviation from actual values at 4 months and 2 years after surgery. However, as there were no significant differences in echocardiographic follow-up days between groups, this limitation might have little effect on the difference in changes of echocardiographic variables between groups. There were 12 (2.2%) patients who did not receive CABG and 26 (4.8%) patients of crossover between CABG+SVR and CABG alone groups, but when the analysis was repeated based on as-treated, the results did not change significantly from the current analysis in terms of changes of SI, grade of mitral regurgitation and diastolic dysfunction and effect of SI on overall survival (data not shown). Follow-up measurement of SI especially after SVR could be very challenging since the original LV apex was cut off after SVR. Therefore there were possibilities of new LV apex being different from original one, which might result in over- or underestimation of LV long-axis dimension and SI. Finally, although every surgeon in the trial was trained for SVR according to the protocol to standardize surgical technique, actual operative technique might be different from centers to centers, which might have some differential effect on the change of LV geometry. Despite these limitations, our results provide an insight into geometric and hemodynamic alteration after CABG+SVR in terms of the impact of the baseline SI on clinical outcome and the impact of SVR compared to CABG alone on subsequent SI as well as other echocardiographic parameters.
Conclusions
This analysis of the STICH trial shows that higher baseline SI was associated with worse survival in patients with ischemic cardiomyopathy and anterior wall akinesia/dyskinesia, undergoing surgical revascularization, whether or not SVR was performed. Although SVR was designed to improve LV geometry, SI worsened after SVR despite improved LVEF and a smaller LV volume. The worsening of SI with SVR was accompanied by worsening of diastolic function and less improvement in mitral regurgitation.
Supplementary Material
Acknowledgments
Funding: This work was supported by the National Heart, Lung and Blood Institute.
Footnotes
Conflicts of interest: none declared.
References
- 1.White H, Norris R, Brown M, Brandt P, Whitlock R, Wild C. Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation. 1987;76:44–51. doi: 10.1161/01.cir.76.1.44. [DOI] [PubMed] [Google Scholar]
- 2.Wong SP, French JK, Lydon A-M, Manda SOM, Gao W, Ashton NG, White HD. Relation of left ventricular sphericity to 10-year survival after acute myocardial infarction. The American Journal of Cardiology. 2004;94:1270–1275. doi: 10.1016/j.amjcard.2004.07.110. [DOI] [PubMed] [Google Scholar]
- 3.Athanasuleas CL, Stanley AWH, Jr, Buckberg GD, Dor V, DiDonato M, Blackstone EH. Surgical anterior ventricular endocardial restoration (SAVER) in the dilated remodeled ventricle after anterior myocardial infarction. Journal of the American College of Cardiology. 2001;37:1199–1209. doi: 10.1016/s0735-1097(01)01119-6. [DOI] [PubMed] [Google Scholar]
- 4.Jones RH, Velazquez EJ, Michler RE, Sopko G, Oh JK, O’Connor CM, Hill JA, Menicanti L, Sadowski Z, Desvigne-Nickens P, Rouleau J-L, Lee KL. Coronary Bypass Surgery with or without Surgical Ventricular Reconstruction. New England Journal of Medicine. 2009;360:1705–1717. doi: 10.1056/NEJMoa0900559. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Pocar M, Di Mauro A, Passolunghi D, Moneta A, Alsheraei AMTA, Bregasi A, Mattioli R, Donatelli F. Predictors of adverse events after surgical ventricular restoration for advanced ischaemic cardiomyopathy. European Journal of Cardio-Thoracic Surgery. 2010;37:1093–1100. doi: 10.1016/j.ejcts.2009.12.006. [DOI] [PubMed] [Google Scholar]
- 6.Oh JK, Velazquez EJ, Menicanti L, Pohost GM, Bonow RO, Lin G, Hellkamp AS, Ferrazzi P, Wos S, Rao V, Berman D, Bochenek A, Cherniavsky A, Rogowski J, Rouleau JL, Lee KL. Influence of baseline left ventricular function on the clinical outcome of surgical ventricular reconstruction in patients with ischaemic cardiomyopathy. Eur Heart J. 2013;34:39–47. doi: 10.1093/eurheartj/ehs021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Velazquez EJ, Lee KL, O’Connor CM, Oh JK, Bonow RO, Pohost GM, Feldman AM, Mark DB, Panza JA, Sopko G, Rouleau JL, Jones RH. The rationale and design of the Surgical Treatment for Ischemic Heart Failure (STICH) trial. J Thorac Cardiovasc Surg. 2007;134:1540–1547. doi: 10.1016/j.jtcvs.2007.05.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr. 2005;18:1440–1463. doi: 10.1016/j.echo.2005.10.005. [DOI] [PubMed] [Google Scholar]
- 9.Oh JK, Pellikka PA, Panza JA, Biernat J, Attisano T, Manahan BG, Wiste HJ, Lin G, Lee K, Miller FA, Jr, Stevens S, Sopko G, She L, Velazquez EJ. Core lab analysis of baseline echocardiographic studies in the STICH trial and recommendation for use of echocardiography in future clinical trials. J Am Soc Echocardiogr. 2012;25:327–336. doi: 10.1016/j.echo.2011.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eilers PHC, Marx BD. Flexible smoothing with B-splines and penalties. Statist Sci. 1996;11:89–102. [Google Scholar]
- 11.Athanasuleas CL, Buckberg GD, Stanley AW, Siler W, Dor V, Di Donato M, Menicanti L, Almeida de Oliveira S, Beyersdorf F, Kron IL, Suma H, Kouchoukos NT, Moore W, McCarthy PM, Oz MC, Fontan F, Scott ML, Accola KA. Surgical ventricular restoration in the treatment of congestive heart failure due to post-infarction ventricular dilation. J Am Coll Cardiol. 2004;44:1439–1445. doi: 10.1016/j.jacc.2004.07.017. [DOI] [PubMed] [Google Scholar]
- 12.Castelvecchio S, Menicanti L, Ranucci M, Di Donato M. Impact of surgical ventricular restoration on diastolic function: implications of shape and residual ventricular size. Ann Thorac Surg. 2008;86:1849–1854. doi: 10.1016/j.athoracsur.2008.08.010. [DOI] [PubMed] [Google Scholar]
- 13.Di Donato M, Castelvecchio S, Burkhoff D, Frigiola A, Raweh A, Menicanti L. Baseline left ventricular volume and shape as determinants of reverse remodeling induced by surgical ventricular reconstruction. Ann Thorac Surg. 2011;92:1565–1571. doi: 10.1016/j.athoracsur.2011.04.062. [DOI] [PubMed] [Google Scholar]
- 14.Di Donato M, Menicanti L, Ranucci M, Castelvecchio S, de Vincentiis C, Salvia J, Yussuf T. Effects of surgical ventricular reconstruction on diastolic function at midterm follow-up. J Thorac Cardiovasc Surg. 2010;140:285–291. doi: 10.1016/j.jtcvs.2009.10.027. e281. [DOI] [PubMed] [Google Scholar]
- 15.Witkowski TG, ten Brinke EA, Delgado V, Ng AC, Bertini M, Marsan NA, Ewe SH, Auger D, Yiu KH, Braun J, Klein P, Steendijk P, Versteegh MI, Klautz RJ, Bax JJ. Surgical ventricular restoration for patients with ischemic heart failure: determinants of two-year survival. Ann Thorac Surg. 2011;91:491–498. doi: 10.1016/j.athoracsur.2010.09.074. [DOI] [PubMed] [Google Scholar]
- 16.Ueno T, Sakata R, Iguro Y, Yamamoto H, Ueno M, Ueno T, Matsumoto K. Mid-term changes of left ventricular geometry and function after Dor, SAVE, and Overlapping procedures. Eur J Cardiothorac Surg. 2007;32:52–57. doi: 10.1016/j.ejcts.2007.02.030. [DOI] [PubMed] [Google Scholar]
- 17.Choi JO, Cho SJ, Yang JH, Kim JH, Song BK, Park YW, Choi SM, Lee SY, Lee SC, Park SW. Systolic long axis function of the left ventricle, as assessed by 2-D strain, is reduced in the patients who have diastolic dysfunction and a normal ejection fraction. Korean Circ J. 2008;38:250–256. [Google Scholar]
- 18.Perloff JK, Roberts WC. The mitral apparatus. Functional anatomy of mitral regurgitation. Circulation. 1972;46:227–239. doi: 10.1161/01.cir.46.2.227. [DOI] [PubMed] [Google Scholar]
- 19.Mitchell GF, Lamas GA, Vaughan DE, Pfeffer MA. Left ventricular remodeling in the year after first anterior myocardial infarction: a quantitative analysis of contractile segment lengths and ventricular shape. J Am Coll Cardiol. 1992;19:1136–1144. doi: 10.1016/0735-1097(92)90314-d. [DOI] [PubMed] [Google Scholar]
- 20.van den Heuvel AF, van Veldhuisen DJ, van der Wall EE, Blanksma PK, Siebelink HM, Vaalburg WM, van Gilst WH, Crijns HJ. Regional myocardial blood flow reserve impairment and metabolic changes suggesting myocardial ischemia in patients with idiopathic dilated cardiomyopathy. J Am Coll Cardiol. 2000;35:19–28. doi: 10.1016/s0735-1097(99)00499-4. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.