Author + information
- Received July 21, 1997
- Revision received May 18, 1998
- Accepted June 2, 1998
- Published online October 1, 1998.
- William D Denney, MD∗,
- W.Evans Kemp Jr., MD∗,* (, )
- Lowell B Anthony, MD‡,
- John A Oates, MD† and
- Benjamin F Byrd III, MD, FACC∗
- ↵*Address for correspondence: Dr. W. Evans Kemp, Jr., 315 MRB II, 2220 Pierce Ave., Vanderbilt University Medical Center, Nashville, Tennessee 37232
Objectives. To study the applicability of a newly developed echocardiographic scoring system in the assessment of carcinoid valvular heart disease.
Background. We investigated prospectively the development, progression and regression of carcinoid valvular heart disease in patients with carcinoid syndrome by serial echocardiography, correlating these features with urinary 5-HIAA levels and clinical data collected during therapy with somatostatin analog.
Methods. Twenty-three patients with carcinoid syndrome underwent serial echocardiographic examinations. An echocardiographic carcinoid valvular heart disease (CVHD) % score was determined from points assigned for tricuspid and pulmonary valve structure and function.
Results. Fifteen patients had no CVHD at study entry (group 1), while 8 patients had findings of CVHD (group 2). Five patients in group 1 developed new CVHD (1B), while one demonstrated progression of CVHD (2B). The remaining patients did not develop (1A) or had no progression of CVHD (2B). Despite major declines in 5-HIAA levels during therapy in most patients, CVHD did not regress. There were significantly lower levels of median baseline 5-HIAA (98.8 vs. 256 mg/24 h), posttreatment 5-HIAA (50.3 vs. 324 mg/24 h) and posttreatment 5-HIAA time integral (37.3 vs. 192 g/24 h∗ days) in group A vs. B (p < 0.05). However, only posttreatment 5-HIAA levels independently predicted the development or progression of CVHD by multiple step-wise regression analysis (p < 0.005), with a threshold observed in the 100 mg/24 h range.
Conclusions. We designed a new echocardiographic scoring system to evaluate CVHD. Correlating echocardiographic scores with biochemical and clinical markers showed that only posttreatment 5-HIAA levels independently predicted the development or progression of CVHD. This study strengthens the association between serotonin secretion and CVHD, as well as introducing a new technique for serial follow-up of these patients.
Carcinoid tumors are uncommon and arise from the enterochromaffin cells of the gastrointestinal tract and bronchi. Carcinoid tumors may secrete large amounts of vasoactive substances, including 5-hydroxytyptamine (serotonin) which is subsequently metabolized to 5-HIAA by the endothelium of the pulmonary and hepatic vasculature. The carcinoid syndrome occurs in less than 4% of patients with carcinoid tumor (1,2)and is generally associated with hepatic metastases and elevated urinary 5-HIAA levels. Clinically, carcinoid syndrome is characterized by flushing, diarrhea, bronchoconstriction, telangiectasias and right heart disease.
Heart disease characteristic of carcinoid has been previously reported in 19% to 66% of patients with carcinoid syndrome (3–6). The characteristic cardiac lesion of carcinoid heart disease consists of fibrous plaques covering normal endothelium of the chambers and valves of the right heart (7). Left heart involvement is uncommon and is generally associated with bronchial carcinoid or right-to-left intracardiac shunting. Although the etiologic agent responsible for carcinoid heart disease has not been identified, several studies have implicated tumor secretory activity by demonstrating significantly higher urinary 5-HIAA levels in patients with carcinoid syndrome who have carcinoid heart disease (4,6).
Several analogs of the naturally occurring peptide somatostatin have been shown to improve diarrhea and flushing in patients with carcinoid syndrome while simultaneously lowering urinary 5-HIAA levels. However, the impact of treatment with these analogs on carcinoid tumor progression and cardiac involvement has yet to be determined.
The present study was designed to investigate prospectively the progression or regression of carcinoid heart disease in patients with carcinoid syndrome using serial 2-dimensional and Doppler echocardiograhpic studies employing a newly developed scoring system. To establish the association of carcinoid heart disease with urinary 5-HIAA levels, we correlated our results with serial urinary 5-HIAA levels and clinical data collected during therapy with somatostatin analog.
From November 1986 to August 1991, 23 patients with carcinoid syndrome were referred to Vanderbilt University for treatment with somatostatin analog. All 23 patients had histologically proven carcinoid, at least one of the typical symptoms of carcinoid syndrome (flushing, diarrhea or bronchoconstriction), and elevated urinary 5-HIAA levels. The mean age at entry was 59 ± 12 years (range 36 to 78 years). There were 14 men and 9 women. The age at entry, duration of carcinoid symptoms and extent of hepatic involvement with tumor were noted at the time of entry into the study. The extent of hepatic involvement with tumor was defined by abdominal computer tomography (CT) scan as no involvement, single hepatic metastasis or multiple hepatic metastases. All patients were begun on somatostatin analog using a protocol designed to determine the maximal tolerated dose. Doses were increased on a monthly basis to the maximal tolerated dose or a predetermined maximal dose (whichever was less). Twenty patients received octreotide while three received lantreotide.
Two-dimensional (2D) echocardiography, pulsed and continuous wave Doppler, and color flow Doppler studies were performed using standard techniques on each patient at entry into the study and serially approximately every 6 months. All studies were performed on Hewlett-Packard Sonos 500 or 1000 instruments and recorded on in videotape. A total of 126 echocardiographic studies were performed during the study.
To facilitate the detection of serial changes in carcinoid heart disease, we devised a scoring system (Fig. 1)which expressed the severity of carcinoid-induced valvular stenosis and insufficiency in the right heart by combined two-dimensional echocardiographic and Doppler criteria. This Carcinoid Valvular Heart Disease (CVHD) score was determined by 1) tricuspid valve appearance, 2) tricuspid regurgitation severity by either spectral pulsed wave or color Doppler flow mapping; 3) pulmonary stenosis severity by spectral pulsed or continuous wave Doppler, and 4) pulmonary insufficiency severity by color Doppler.
Tricuspid valve appearance was assigned 4 points and was graded on a 1 to 4 point scale as normal, mildly, moderately, or severely diseased. Grading was based upon qualitative assessment of leaflet thickness, leaflet mobility and leaflet retraction using all available echocardiographic views. Tricuspid valves with normal appearance received only 1 point while severely affected valves which were thickened, fixed and retracted received 4 points.
Tricuspid regurgitation was also assigned 4 points in the CVHD score and was graded based upon the color Doppler appearance of the regurgitant jet (or pulsed wave Doppler mapping in a few of the earliest studies) using a method similar to that reported by Suzuki (8). The severity of TR was graded from 1 to 4 by measuring the ratio of the color Doppler regurgitant jet “mosaic” area to the right atrial RA area in the apical 4 chamber view (TR/RA area ratio <10% = 1 point, 10–20% = 2 points, 20–40% = 3 points, and >40% = 4 points). When pulsed wave Doppler mapping was used, the severity of regurgitation was quantitated using the length of the regurgitant jet (8).
Pulmonic valve stenosis (PS) was assigned 4 points as well by measuring the peak Doppler velocity at the pulmonary valve. One point was scored for peak velocities <2 m/s (gradient <16 mm Hg), 2 points for 2–2.5 m/s (gradient 16–25 mm Hg). Higher severity scores were based on peak Doppler velocities and the severity of TR (see Fig. 1).
Pulmonic valve insufficiency (PI) was assigned only 2 points in the CVHD scoring system since no method for assessing its severity is well validated and its clinical significance in carcinoid heart disease is relatively small. PI was graded using a color Doppler method analogous to the one used for aortic insufficiency (9). Using the parasternal short axis view at the aortic valve level, we determined the ratio of the width of the regurgitant jet neck to the width of the pulmonary valve annulus. One point was scored for ratios less than 0.33, and for ratios greater than 0.33 2 points were scored.
The maximum number of possible points for any carcinoid heart disease echo study was thus 14. However, since some studies were inevitably too technically difficult to grade in one or two categories, the CVHD score was reported as a percentage of the points possible on that study. The CVHD% score thus equaled the total points scored divided by the points possible on that study, expressed as a percentage. The maximum normal score equaled 4 of 14 possible points (CVHD% score < 30), as in a patient with mild TR on color Doppler study, a structurally normal tricuspid valve, and a normal pulmonary valve.
All the echo studies were scored by a single observer unaware of the patients’ previous echo scores or clinical and laboratory data. Fifteen randomly selected echo studies were rescored to determine the intraobserver variability. Based upon this result, a significant change in score was defined as greater than twice the standard deviation of the mean intraobserver variability. Patients were then grouped based upon the stability or progression of their carcinoid heart disease by comparing their CVHD% scores from initial and final echo studies.
Urinary 5-HIAA excretion was determined by an assay utilizing fluorescence polarization immunoassay technology (TDx Assay, Abbott Laboratories) on samples obtained before December 31, 1992 and by an assay utilizing high performance liquid chromatography on samples obtained after this date. Multiple 24-h urine collections were obtained on each patient as they were admitted for serial reevaluation of their response to treatment with somatostatin analog at a minimum of 6 month intervals. Mean pre- and posttreatment 24 h urinary 5-HIAA levels were determined for each patient. Mean posttreatment 5-HIAA values were determined as the sum of all 24-h urine values, obtained after therapy with somatostatin analog was initiated, divided by the number of days sampled. Mean pre- and posttreatment 24-h urinary 5-HIAA levels were expressed as mg/24 h. As a measure of cumulative exposure to serotonin for each patient, time-integrated urinary 5-HIAA levels were determined (Fig. 2)using the area under the curve of the multiple 24-h urinary 5-HIAA levels obtained over the period of follow up (result expressed in grams/24 h × days).
Many of the variables were not normally distributed after grouping, so the median and interquartile ranges were used as measures of central tendency and variation. The Mann-Whitney rank sum Utest was used to compare the data from groups of patients who were either stable or progressed with respect to echocardiographic evidence of carcinoid valvular heart disease. Univariate and multiple step-wise regression analyses were used to determine the clinical and laboratory predictors of echocardiographic progression of carcinoid valvular heart disease. Statistical significance was determined at the p < 0.05 level.
Typical right heart carcinoid valvular abnormalities were observed in 8 patients (35%) on their initial 2D echocardiogram and Doppler study, while the remaining 15 patients showed no evidence of carcinoid heart disease. Those patients without carcinoid heart disease on the initial 2D echo were placed in group 1; those with carcinoid heart disease were placed in group 2. Since the intraobserver reproducibility study of 15 patients revealed a mean absolute difference in CVHD% score of 3.8 with a standard deviation of 4.7, a >10 point change in score was judged to be significant. As shown in Table 1, patients were thus further classified into those who demonstrated no echocardiographic progression or development of carcinoid heart disease (subgroup A) and those whose carcinoid heart disease progressed or developed (subgroup B).
Of the 15 patients who entered the study without evidence of carcinoid heart disease (group 1, median CVHD% score 29), 10 patients remained unaffected (group 1A, median absolute increase in CVHD% score of 0), while the other 5 developed new echocardiographic changes typical of carcinoid heart disease (group 1B, median absolute increase in CVHD% score of 28). Of the 8 patients demonstrating echocardiographic evidence of carcinoid heart disease at entry (group 2, median CVHD% score of 70), 7 patients remained stable (group 2A, median absolute increase in CVHD% score of 4) while 1 patient demonstrated progressive carcinoid heart disease (group 2B, median increase in CVHD% score of 15). Echocardiographic regression of carcinoid heart disease was not observed in any patient despite treatment with somatostatin analog. That is, an improvement in CVHD% score of 10 or more was never sustained on more than 2 serial echocardiograms.
Comparison of clinical variables at presentation in patients without echocardiographic evidence of carcinoid heart disease (group 1) and those with carcinoid heart disease (group 2) is summarized in Table 2. All but 2 patients showed either single or multiple hepatic metastases on abdominal CT scan. Although patients with carcinoid heart disease at entry tended to have carcinoid symptoms of greater duration, higher baseline urinary 5-HIAA levels and multiple metastases, these differences did not reach statistical significance.
Clinical variables from all patients who showed no significant echocardiographic evidence of progression/development of carcinoid valvular heart disease (group A) were then compared with those who demonstrated progression/development (group B). As shown in Figure 3, there were statistically significant differences in baseline urinary 5-HIAA levels (98.8 [59–186] vs. 256 [145–356] mg/24 h), posttreatment urinary 5-HIAA levels (50.3 [27–90] vs. 324 [103–458] mg/24 h), and posttreatment urinary 5-HIAA integral (37.3 [21–79.2] vs. 192 [59–703] g/24 h × days) (median [range]) (p < 0.05). There was no significant difference between these groups in the duration of carcinoid symptoms prior to entry, duration of treatment with somatostatin analogue, or percent change in urinary 5-HIAA levels (Table 3).
Table 4compares clinical variables from the subgroup of patients who entered the study without echocardiographic evidence of carcinoid heart disease and remained stable (group 1A) with those who developed new evidence of disease (group 1B). Patients from group 1B had significantly higher baseline urinary 5-HIAA levels (319 [193–356] vs. 77 [42–112] mg/24 h), post-treatment urinary 5-HIAA levels (442 [207–458] vs. 57.8 [20–90] mg/24 h), and posttreatment urinary 5-HIAA integrals (238 [59–703] vs. 34.2 [12–79] g/24 h∗ days) (median [range]) (p < 0.05) (Fig. 4). No statistically significant differences were shown in the duration of carcinoid symptoms prior to entry, percent change in urinary 5-HIAA levels, or duration of therapy with somatostatin analogue.
As shown in Table 5, three clinical variables were found to predict carcinoid valvular heart disease progression by univariate analysis. These were the posttreatment urinary 5-HIAA levels, the baseline urinary 5-HIAA levels and the 5-HIAA integral (p < 0.05). However, multiple step-wise regression analysis showed that only the posttreatment urinary 5-HIAA levels independently predicted the progression of carcinoid valvular heart disease (p < 0.005).
During the 6-year-study period, the mortality rates were 3/10 (30%) for group 1A, 3/5 (60%) for group 1B, 1/7 (14%) for group 2A and 1/1 (100%) for group 2B.
This prospective study used a new echocardiographic scoring system to document progression or regression of carcinoid valvular heart disease in serial studies. This CVHD scoring system is based upon the structure of the right heart valves as shown by 2D echocardiography, and upon their function as shown by Doppler echocardiography. In this study of patients with histologically proven metastatic carcinoid tumor, the 35% incidence of cardiac involvement on the initial echocardiogram is similar to that of previous studies (4–6). Although patients with cardiac involvement on their initial echocardiogram tended to have had symptoms of longer duration, higher baseline urinary 5-HIAA levels, and more frequent multiple hepatic metastases, none of these differences reached statistical significance. Pellikka et al. in an echocardiograhic study of 132 patients with carcinoid syndrome noted no difference in patients with carcinoid heart disease and those without carcinoid heart disease with regard to symptom duration, time from diagnosis or location of the primary tumor (10). However, they did note higher mean pretreatment 5-HIAA levels in patients with carcinoid heart disease than in those without carcinoid heart disease. The ranges of values were quite broad with moderate overlap between the 2 groups.
In this study, the development and progression of carcinoid heart disease during treatment with somatostatin analog were clearly linked to elevated urinary 5-HIAA levels. The 5 patients who developed carcinoid heart disease de novo during the study had significantly higher median urinary 5-HIAA levels and higher urinary 5-HIAA integrals than those who remained free of carcinoid heart disease. None of the other clinical variables, including duration of therapy, differed significantly between those who suffered progressive carcinoid heart disease and those who did not.
Our scoring system did not include mitral CVHD, which is only seen in 7% of cases (10). Only 1 patient in this study exhibited mitral valve disease with thickening and decreased excursion of the mitral valve leaflets (4%). This patient had a bronchial carcinoid tumor and simultaneously showed progression of pre-existing right heart involvement.
Stabilization or regression
Interestingly, regression of carcinoid disease (defined as a significant decrease in CVHD% score of >10 points on serial echocardiography) was not observed during treatment with somatostatin analog. Although the median posttreatment urinary 5-HIAA level was reduced to less than 50% of pretreatment level in 6 patients who entered the study with carcinoid heart disease, CVHD% score by echocardiography did not change significantly in these patients. There was a trend toward lower posttherapy 5-HIAA levels in the group with stable carcinoid heart disease when compared with the group with progressive or new carcinoid heart disease. Failure of this trend to reach statistical significance may be due to the small sample size as well as significant overlap of the ranges for percent change in 5-HIAA levels. Nevertheless, if more patients with pre-existing carcinoid heart disease can be followed after major reduction in urinary 5-HIAA levels, effective therapy may be shown to stabilize or even regress carcinoid disease.
Although it has been hypothesized that exposure of the endocardium to elevated levels of serotonin (4,6)or other secreted tachykinins such as neuropeptide K or substance P(4) might lead to fibroblast proliferation, the exact etiology of the fibrous plaques is obscure. Of the three clinical variables identified in the univariate analysis as predictors of progression of carcinoid heart disease (baseline urinary 5-HIAA levels, posttreatment urinary 5-HIAA levels, and the posttreatment urinary 5-HIAA time integral), only the posttreatment urinary 5-HIAA level was significant by multiple step-wise regression analysis. Failure of the posttreatment urinary 5-HIAA time integral to independently predict progression of carcinoid heart disease suggests that the absolute elevation in serotonin secretion is more important than cumulative exposure to serotonin. That is, there may be a threshold value of serotonin secretion above which fibroblast proliferation is activated. As shown in Figures 3 and 4, the threshold for urinary 5-HIAA values during therapy appears to be in the 100 mg/24 h range.
Morbidity and mortality
Since right heart failure contributes significantly to the morbidity and mortality associated with malignant carcinoid tumors, predictors of carcinoid heart disease development may be useful. Indeed, mortality in the patients who developed new (group 1B) or progressive (group 2B) CVHD was more than twice that of patients with no (group 1A) or stable (group 2A) carcinoid heart disease. Since persistent elevations in urinary 5-HIAA levels despite therapy (>100 mg/24 h) were independently associated with development or progression of carcinoid heart disease, aggressive attempts to reduce 5-HIAA levels below this threshold appear justified.
The CVHD% score provides a compilation of right heart pathology which can be used to monitor CVHD development and progression quantitatively. A larger study is needed to determine whether aggressive reduction of urinary 5-HIAA levels will reliably prevent or halt the progression of cardiac involvement. It has been suggested that a lower 5-HIAA level prior to valvular surgery correlates with improved postoperative survival (11). Surgical intervention with valvuloplasty or valve replacement remains the best means of improving right heart valvular function since regression of carcinoid valvular heart disease was not observed in this study during treatment with somatostatin analog.
Limitations and conclusions
One limiting feature of this study is the relatively small number of patients following during therapy for this relatively uncommon condition, metastatic carcinoid tumor. Although the differences in 5-HIAA levels were statistically significant, we still can only speculate on the role serotonin plays in carcinoid heart disease. Nevertheless, this prospective study demonstrated the usefulness of a new echocardiographicscoring system to evaluate the presence and severity of carcinoid heart disease by serial 2D and Doppler echocardiography. In those patients demonstrating echocardiographic progression of carcinoid heart disease, urinary 5-HIAA levels were significantly higher both before and during therapy with somatostatin analog. This study strengthens the association between serotonin secretion and carcinoid heart disease, and it also demonstrates the application of serial echocardiography in patients with malignant carcinoid tumors.
- 5-hydroxyindoleacetic acid
- carcinoid valvular heart disease
- computed tomography
- tricuspid regurgitation
- pulmonic valve insufficiency
- pulmonic valve stenosis
- right atrium
- Received July 21, 1997.
- Revision received May 18, 1998.
- Accepted June 2, 1998.
- American College of Cardiology
- Lundin L,
- Norheim I,
- Landelius J,
- et al.
- Suzuki Y,
- Kambara H,
- Kadota K,
- et al.
- Perry G.J,
- Helmcke F,
- Nanda N,
- et al.
- Pellikka P.A,
- Tajik J.A,
- Khandheria B.K,
- et al.
- Connoly H.M,
- Nishimura R.A,
- Smith H.C,
- et al.