Cardiac fibrosis is an important complication of intestinal carcinoid disease, with resulting valvular and ventricular dysfunction due to endocardial fibrosis. Evaluation of right ventricular (RV) function in these patients has focused on valvular involvement. The aim of this study was to investigate whether myocardial strain by echocardiography can detect RV dysfunction in patients with carcinoid disease.
Eighty-nine patients with intestinal carcinoid and 50 healthy individuals were studied. Strain measurements were assessed by speckle-tracking echocardiography. The average of the three lateral RV segments was calculated as RV strain. Left ventricular global strain was calculated from a 16-segment model.
Carcinoid heart disease was present in 15 of the 89 patients. RV strain was reduced in patients with carcinoid disease compared with healthy controls (−20.6 ± 5.0% vs −26.9 ± 4.4%, P < .001). RV function by strain was not significantly different in patients with and without carcinoid heart disease (−21.2 ± 5.7% vs −20.5 ± 4.8%, P = .59). Excluding patients with overt carcinoid heart disease, RV strain was reduced in patients with 5-hydroxyindoleacetic acid levels above the normal range compared with those with levels within the normal range (−19.4 ± 5.4 vs −21.6 ± 3.7%, P = .05).
RV function by myocardial strain was reduced in patients with carcinoid disease independently of valvular involvement. This indicates that myocardial strain by echocardiography provides added information about RV function in patients with intestinal carcinoid disease.
Fibrosis is an important complication of neuroendocrine tumors and is most often associated with small intestinal, appendiceal, and proximal colonic neuroendocrine tumors (midgut carcinoid tumors). The development of endocardial fibrosis, known as carcinoid heart disease (HD), is of particular clinical concern and leads to increased morbidity and mortality in these patients. Carcinoid HD occurs in 20% to 70% of patients with metastatic carcinoid tumors and results in retraction and fixation of the heart valves, sometimes necessitating surgery. Currently, there is no clinical method to predict and determine the development of fibrosis, especially at an early stage. The etiology behind endocardial fibrosis is not fully understood, but it involves vasoactive substances (e.g., serotonin) released from the tumor and affects the heart through poorly understood mechanisms. Mainly the valves in the right heart are affected, likely because of the inactivation of tumor products in the pulmonary circulation.
The evaluation of cardiac involvement in patients with carcinoid disease has so far been limited to estimations of valvular regurgitation and/or stenosis, with tricuspid regurgitation observed most frequently. However, because carcinoid HD not only affects valvular but also mural endocardium, assessment of right ventricular (RV) function as a measure of mural endocardial fibrosis may be of clinical importance in these patients. Methods of evaluating RV function are limited by the complex geometry and load dependency of the right ventricle. Novel echocardiographic techniques have been shown to improve the assessment of cardiac function. Myocardial strain by echocardiography can accurately quantify regional myocardial function, including both the left and right ventricles. To date, however, there is paucity of data investigating RV mural endocardial fibrosis and its functional characteristics in patients with carcinoid disease.
Left-sided valvular pathology occurs in approximately 10% of patients with carcinoid HD and is associated with right-to-left shunting, bronchial carcinoid, or poorly controlled carcinoid syndrome. Although some studies have examined left ventricular (LV) function using traditional means, to our knowledge, studies assessing LV function using myocardial strain have not yet been performed in this patient group.
We hypothesized that myocardial strain can detect reduced RV and LV function in patients with carcinoid disease either concomitant with or without valvular involvement and might therefore be a complementary marker of carcinoid HD.
A total of 89 consecutive patients with histologically verified small intestinal ( n = 86), appendiceal ( n = 2), or proximal colonic ( n = 1) carcinoid tumors were included in this study between 2006 and 2007. The inclusion criterion was a diagnosis of midgut carcinoid disease (by definition carcinoid tumors arising from the jejunum, ileum, appendix, or proximal colon) with histologic confirmation from either the primary tumor or, if the patient had not undergone resection of the primary tumor, from biopsy of a metastasis. A diagnosis of midgut carcinoid was confirmed via review of all original biopsy reports, including documentation of staining for markers of neuroendocrine tumors such as synaptophysin, chromogranin A, and neuron-specific enolase. In those patients in whom the primary tumor was not resected, localization of the primary tumor was dependent on radiologic findings. No patients with pulmonary carcinoid tumors were included. Carcinoid syndrome was defined as the presence of flushing and/or diarrhea and was present in 81% of the patients.
Patients affected by another fibrotic disease such as systemic sclerosis, those aged <18 years, and those unable to give informed consent were excluded from the study. None of the patients had suspected coronary artery disease on the basis of medical history and/or electrocardiographic changes. The mean time interval from the diagnosis of midgut carcinoid tumor to inclusion was 5.2 ± 5.1 years. Eighteen of the 89 patients had previously undergone radical resection of their tumors. Of the remaining 71 patients, 62 had liver metastases at the time of inclusion. All of the patients routinely underwent computed tomography of the abdomen as part of their admission. To assess tumor burden, reanalysis of the abdominal computed tomographic scans was performed to determine the number of liver metastases for each patient.
A 24-hour urine collection for 5-hydroxyindoleacetic acid (5-HIAA; normal range, 0.8–3.8 μmol/mmol creatinine) was obtained in 86 of the 89 patients.
Fifty healthy individuals were included from the hospital staff and their families for comparison. The results of clinical examinations, electrocardiography, and echocardiography were normal for all control subjects.
All participants underwent complete transthoracic echocardiography, including two-dimensional, color flow, and spectral Doppler as well as Doppler tissue imaging using a Vivid 7 machine (GE Vingmed Ultrasound AS, Horten, Norway). Offline tissue Doppler analysis was performed using commercially available software (EchoPAC; GE Vingmed Ultrasound AS). LV ejection fraction was calculated using Simpson’s method. Carcinoid HD was defined as a combination of right-sided valvular dysfunction (i.e., more than mild regurgitation and/or any valvular stenosis) and pathologic morphology (i.e., valve leaflet thickening, shortening, retraction, hypomobility, or incomplete coaptation). Tricuspid valve regurgitation was assessed using current guidelines and graded as none (0), mild (1), moderate (2), or severe (3). Tricuspid annular plane systolic excursion (TAPSE) was measured by M-mode echocardiography at the RV free wall in the apical four-chamber view. RV fractional area change was calculated from RV area at end-systole and RV area at end-diastole as previously described. The RV index of myocardial performance (RIMP) was assessed in all participants.
Two-Dimensional Strain Echocardiography
Two-dimensional grayscale images were acquired in the standard parasternal and apical (apical four-chamber, two-chamber, and long-axis) views at 67 ± 26 frames/sec for patients and 67 ± 11 frames/sec for controls. Three cardiac cycles were recorded. All images were stored digitally for subsequent offline analysis. Myocardial deformation measurements were performed using speckle-tracking echocardiography. RV longitudinal strain was obtained from the apical four-chamber view. Peak systolic strain was assessed in the three lateral RV segments, defined as the lateral basal, lateral mid, and lateral apical segments ( Figure 1 ). These strain values were averaged and defined as average RV strain. The interventricular septum was attributed to LV function. LV strain was obtained from apical four-chamber, two-chamber, and long-axis views. Global LV longitudinal strain was obtained by averaging all segmental values for peak systolic strain in a 16-segment model. In addition, strain measurements from the three LV septal segments from apical four-chamber view were averaged and compared with an average of the rest of the LV segments.
Strain analyses were performed offline by cardiologists blinded to patients’ clinical and biochemical data. However, overt valvular disease was apparent during analyses. Observers were not blinded regarding controls’ status as healthy individuals. Strain parameters could be assessed in 94% of the RV segments in the study group and in 95% of the subjects in the control group.
Magnetic Resonance Imaging (MRI)
MRI was performed in nine patients without overt carcinoid HD using 1.5-T units (Magnetom Vision Plus or Magnetom Sonata; Siemens Medical Systems, Erlangen, Germany) and a phased-array body coil. Axial and sagittal T1 turbo spin-echo images, multiple axial, and one sagittal cine loop covering the right and left ventricles were recorded. Delayed-enhancement gadolinium images were acquired to assess myocardial involvement.
This study was approved by the regional ethics committee for research. All individuals gave their written informed consent to participate in the study.
Data are presented as mean ± SD. Comparisons of means were analyzed using analysis of variance with Bonferroni’s post hoc correction for multiple comparisons for normally distributed data. Measurements of 5-HIAA are presented as median (interquartile range) and were analyzed using Mann-Whitney U tests. Proportions were compared using χ 2 and Fisher’s exact tests as appropriate. Linear regression analysis was performed to compare RV and LV myocardial function by strain. Spearman’s rank correlation was used to correlate the number of liver metastases and RV function by strain echocardiography. Reproducibility is expressed as intraclass correlation coefficients for single measures. For all statistical analyses, P values were two sided, with results < .05 considered significant. SPSS version 16.0 (SPSS, Inc., Chicago, IL) was used for statistical analysis.
Table 1 summarizes the major clinical characteristics. Age, sex, and heart rate were similar among patients with carcinoid disease compared with healthy individuals. A total of 74 patients (83%) had no carcinoid HD, while 15 (17%) had carcinoid HD by the current definition ( Table 2 ). The patients with carcinoid HD were older compared with those without (69 ± 19 vs 60 ± 11 years, P = .005). Time since diagnosis was similar in patients with and without carcinoid HD ( P = .50). Residual tumor was present in all 15 patients (100%) with overt carcinoid HD and in 56 of 74 patients (76%) without carcinoid HD ( P = .03). Only four patients had signs of patent foramen ovale ( n = 3) or atrial septal defect ( n = 1) on transthoracic echocardiography. Three of these patients had left-sided valvular involvement with more than moderate regurgitation.
|Variable||Healthy controls ( n = 50)||Patients with carcinoid disease ( n = 89)||P ∗|
|Age (y)||57 ± 21||61 ± 12||.22|
|Women||24 (48%)||44 (49%)||.95|
|Heart rate (beats/min)||65 ± 11||67 ± 13||.71|
Average RV strain from the three lateral segments was lower in the groups of patients with carcinoid disease compared with healthy individuals ( P < .001; Table 2 ). RV strain was not significantly reduced in patients with carcinoid HD compared with those without carcinoid HD ( P = .99). Patients without carcinoid HD had reduced TAPSE compared with healthy controls (23 ± 4 vs 26 ± 3 mm, P = .001), but no differences were observed between patients with and without carcinoid HD ( P = .75). More than mild tricuspid regurgitation was observed in 14 of the 15 patients who fulfilled definitions of carcinoid HD. Ten patients had more than minimal pulmonary regurgitation, and two had pulmonary stenosis.
|Variable||Healthy controls ( n = 50)||Patients without carcinoid HD ( n = 74)||Patients with carcinoid HD ( n = 15)||P ‡|
|Time since diagnosis (y)||5.0 ± 5.3||6.0 ± 4.9||.50|
|LV ejection fraction (%)||65 ± 5||66 ± 7||61 ± 7 †||.04|
|TAPSE (mm)||26 ± 3||23 ± 4 ∗||24 ± 5||.001|
|RV end-systolic area (cm 2 )||9.0 ± 3.3||9.3 ± 3.2||13.0 ± 6.8 ∗ †||.002|
|RV end-diastolic area (cm 2 )||17.7 ± 4.9||17.5 ± 4.7||23.6 ± 10.0 ∗ †||.001|
|RV fractional area change (%)||50 ± 9||47 ± 10||45 ± 13||.20|
|RIMP (%)||37 ± 12||47 ± 16 ∗||44 ± 18||.03|
|Average RV lateral strain (%)||−26.9 ± 4.4||−20.5 ± 4.8 ∗||−21.2 ± 5.7 ∗||<.001|
|LV global strain (%)||−21.3 ± 2.0||−18.4 ± 2.6 ∗||−18.1 ± 2.9 ∗||<.001|
|Average LV septal strain (%)||−20.6 ± 2.4||−18.4 ± 3.2 ∗||−18.9 ± 3.0||<.001|
|LV strain mean rest of left ventricle (%)||−21.6 ± 2.2||−17.6 ± 3.0 ∗||−18.3 ± 3.0 ∗||<.001|
|5-HIAA (μmol/mmol creatinine)||3.8 (1.9–14.6)||20.1 (6.1–46.5)||.004|