Author + information
- Received December 7, 1998
- Revision received July 22, 1999
- Accepted October 18, 1999
- Published online February 1, 2000.
- John P Reilly, MDa,
- Paul A Tunick, MD, FACCa,* (, )
- Robert J Timmermans, MDa,
- Bruce Stein, MDa,
- Barry P Rosenzweig, MD, FACCa and
- Itzhak Kronzon, MD, FACCa
- ↵*Reprint requests and correspondence: Dr. Paul A. Tunick, New York University School of Medicine, 560 First Avenue, New York, NY 10016
The study examined the value of contrast echocardiography in the assessment of left ventricular (LV) wall motion in intensive care unit (ICU) patients.
Echocardiograms done in the ICU are often suboptimal. The most common indication is the evaluation of LV wall motion and ejection fraction (EF).
Transthoracic echocardiograms were done in 70 unselected ICU patients. Wall motion was evaluated on standard echocardiography (SE), harmonic echocardiography (HE), and after intravenous (IV) contrast echocardiography (CE) using a score for each of 16 segments. A confidence score was also given for each segment with each technique (unable to judge; not sure; sure). The EF was estimated visually for each technique, and a confidence score was applied to the EF.
Uninterpretable wall motion was present in 5.4 segments/patient on SE, 4.4 on HE (p = 0.2), and 1.1 on CE (p < 0.0001). An average of 7.8 segments were read with surety on SE, 9.2 on HE (p = 0.1), and 13.7 on CE (p < 0.0001). Ejection fraction was uninterpretable in 23% on SE, 13% on HE (p = 0.14), and 0% on CE (p = 0.002 vs. HE; p < 0.0001 vs. SE). The EF was read with surety in 56% of patients on SE, 62% on HE (p = 0.47), and 91% on CE (p < 0.0001). Thus, wall motion was seen with more confidence on CE. More importantly, the actual readings of segmental wall motion and EF significantly differed using CE.
CE should be used in all ICU patients with suboptimal transthoracic echocardiograms.
Patients in the intensive care unit (ICU) are typically quite ill, and a complete understanding of their hemodynamics is essential. Bedside portable echocardiography in the ICU is an important tool in managing such critically ill patients, providing crucial anatomic and hemodynamic data, and often rendering invasive monitoring unnecessary (1). Left ventricular wall motion and ejection fraction impact upon a patient’s capacity to withstand the hemodynamic demands
of illness and are prognostic indicators of survival. Echocardiographic imaging in the ICU, however, can be limited. Frequently, ICU patients are unable to cooperate with the sonographer, and cannot always be optimally positioned. Mechanical ventilation, bandages, lung disease, subcutaneous emphysema, chest tubes, and poor lighting conditions may all impart additional technical obstacles. Because of this, endocardial resolution is often suboptimal, preventing the accurate assessment of segmental and global wall motion.
Echocardiographic contrast media consisting of albumin sonicated in the presence of inert gases such as octafluoropropane (Optison) are safely administered intravenously, cross the pulmonary vascular bed, and provide excellent opacification of the left heart (2). Within 1 min of their administration there is clear delineation of endocardial borders, facilitating interpretation of left ventricular wall motion. Contrast-enhanced echocardiography has a range of applications including treadmill (3)and dobutamine stress echocardiography (4), the evaluation of myocardial perfusion at rest (5)and during myocardial infarction (6)and thrombolysis (7). It has also proven useful for detection of intrapulmonary shunts in liver disease (8), assessment of endothelial function (9), assessment of myocardial viability (10), evaluation of left atrial appendage stasis (11)and quantification of left to right shunts through atrial septal defects (12). We undertook the current study to assess the benefits of contrast echocardiography for the evaluation of left ventricular function in ICU patients.
Portable, bedside transthoracic echocardiograms were performed in 70 unselected ICU patients between April 14 and June 25, 1998. This group represented 26% of the 273 bedside transthoracic echocardiograms performed on ICU patients during this period. These patients were randomly chosen (there was no selection a priori based on the quality of the standard echocardiogram). The mean age of the patients studied was 69.6 ± 13.4 years (range 26 to 94 years). The study population comprised 40 men (57%) and 30 women. The large majority of these studies (54 echocardiograms, 77%) were requested for the evaluation of left ventricular function. Additional referral indications were pericardial effusion (19 patients, 27%), and valve disease (16 patients, 23%). The patients were in the ICU for a range of critical problems, including postoperative management (19 patients), postmyocardial infarction or unstable angina (22 patients), sepsis (8 patients), respiratory failure/pulmonary edema (7 patients), postcardiac arrest (5 patients) and various other medical illnesses. Twenty-two patients (31%) were supported with mechanical ventilation.
Experienced sonographers performed bedside portable transthoracic echocardiograms in the ICU using a Hewlett-Packard (Andover, Massachusetts) Sonos 5500 echocardiographic system. The sonographers obtained the best possible fundamental-imaging echocardiograms (standard imaging), followed by echocardiograms using second-harmonic imaging. Harmonic imaging was done with both 1.8 and 2.1 frequencies, and the best images obtained were used. Finally, harmonic imaging was repeated after the intravenous (IV) injection of 0.5 to 1.0 cc of Optison (contrast imaging). This contrast agent opacified the cardiac blood pool (Fig. 1). Left ventricular wall motion was determined from standard, harmonic and contrast imaging by the agreement of two experienced echocardiographers. All 70 contrast studies were recorded on a separate tape, and these were interpreted in a blinded fashion on a different day from the standard and harmonic images.
Sixteen left ventricular segments were analyzed in three standard views: parasternal long axis, apical four chamber, and apical two chamber. On the long axis view, the septum and posterior wall were divided into two segments each (basal and mid). On the apical views, each wall (septum and lateral on the four-chamber view; anterior and inferior on the two-chamber view) was divided into three segments (basal, mid, and apical). Wall motion for each segment was graded as normal, hypokinetic (mildly, moderately or severely), akinetic, or dyskinetic.
In addition, a Confidence Score was assigned to reflect the degree of confidence the echocardiographers had in their reading of each segment. Wall segments were read as Confidence Score A (uninterpretable), Confidence Score B (interpretable but not sure) or Confidence Score C (sure). Surety (Confidence Score C) was defined as convincing endocardial edge delineation, allowing definitive evaluation of wall motion. No estimate of segmental wall motion (normal, hypokinetic, and so forth) could be made if that segment’s Confidence Score was A (uninterpretable). If the interpreters believed they could offer a reading, and would have done so clinically, but were not entirely confident (sure) of the reading, Confidence Score B was given. If no interpretation could be given at all, Confidence Score A was given. The “A” segments were visualized to a much lesser degree.
Overall ejection fraction was then determined by visual estimation (13)for each technique (standard, harmonic, and contrast) as increased (>70%), normal (50% to 70%), mildly reduced (40% to 49%), moderately reduced (30% to 39%), or severely reduced (<30%). An ejection fraction Confidence Score of A (uninterpretable), B (interpretable but not sure), or C (sure) was then assigned for each patient. No estimate of the ejection fraction could be made if the Ejection Fraction Confidence Score was A (uninterpretable).
Continuous variables were analyzed using Student’s t-test (paired, two-tailed) (Microsoft Excel, Microsoft Office 98), and discontinuous variables were analyzed using the chi-square test. A p value ≤0.05 was considered to be statistically significant.
Uninterpretable wall motion (Confidence score a)
Using standard imaging, an average of 5.4 of 16 segments per patient (34%) were uninterpretable (Table 1). A similar number of segments (4.4, 28%, p = 0.2) were uninterpretable with harmonic imaging. However, with contrast imaging, only 1.1 segments per patient (7%) were uninterpretable. This was significantly fewer uninterpretable segments per patient than with standard imaging (p < 0.0001) or harmonic imaging (p < 0.0001).
Segments read with surety (Confidence Score C)
Using standard imaging, an average of 7.8 of 16 segments per patient (49%) were read with surety (Table 1). The small increase in the number of segments read with surety with harmonic imaging (9.2, 58%) was not statistically significant (p = 0.1). With contrast imaging, however, 13.7 of 16 segments per patient (86%) were read with surety. This was a statistically significantly increase in reading confidence over that with standard imaging (p < 0.0001) and over that with harmonic imaging (p < 0.0001).
Uninterpretable segments that came to be read with surety (change in Confidence Score from A to C)
With standard imaging, an average of 5.4 segments per patient (34%) were uninterpretable. Harmonic imaging allowed an average of only 0.9 (17%) of these same 5.4 segments per patient to be read with surety (Table 2). This added value of contrast imaging over harmonic imaging (each compared to standard imaging) was statistically significant (p < 0.0001).
Improved Confidence Score (either from A to B, A to C, or B to C)
Harmonic imaging allowed 2.7 segments per patient (17%) to be read with a better confidence score than with standard imaging (Table 2). Contrast echocardiography improved the confidence score of 6.8 segments per patient (43%) compared with standard imaging (Table 2). This added value of contrast echocardiography over harmonic imaging (each compared to standard imaging) was also statistically significant (p < 0.0001).
Different wall motion interpretation
Interpretation of wall motion (normal, three grades of hypokinesis, akinesis, or dyskinesis) with harmonic imaging was different by at least one grade in 2.2 segments per patient (14%) compared to standard imaging (Table 3). Contrast echocardiography changed the interpretation of wall motion in 7.1 segments per patient (44%) compared to standard imaging (Table 3). Contrast echocardiography led to a change in wall motion interpretation significantly more frequently than did harmonic imaging, when each was compared to standard imaging (p < 0.0001).
Uninterpretable ejection fraction (Confidence Score A)
The readers were unable to evaluate the ejection fraction in 16 patients (23%) using standard imaging, in 9 patients (13%) using harmonic imaging, and in no patients (0%) using contrast imaging (Table 1). There was no significant difference between standard and harmonic imaging (p = 0.14). However, there were significantly fewer uninterpretable ejection fractions on contrast imaging than on either standard imaging (p < 0.0001) or harmonic imaging (p = 0.002).
Ejection fraction read with surety (Confidence Score C)
The readers were able to evaluate the ejection fraction with surety in 39 patients (56%) using standard imaging, in 42 patients (62%) using harmonic imaging, and in 64 patients (91%) using contrast imaging (Table 1). The number of studies read with surety in evaluating the ejection fraction was therefore significantly greater for contrast imaging than for either standard imaging (p < 0.0001) or for harmonic imaging (p < 0.0001). There was no significant difference between harmonic imaging and standard imaging in this regard (p = 0.47).
Uninterpretable ejection fraction that came to be read with surety (change in Confidence Score from A to C)
There were 16 patients whose ejection fractions were uninterpretable on standard imaging. Harmonic imaging allowed 3 of these 16 patients (19%) to have their ejection fractions read with surety (Table 4). Contrast imaging permitted 11 of these 16 patients (69%) to have their ejection fractions read with surety (Table 4). This added value of contrast imaging over harmonic imaging was statistically significant (p = 0.004). In addition, contrast imaging allowed for the estimation of ejection fraction with surety in 4 of the 9 patients (44%) whose ejection fractions had been uninterpretable by harmonic imaging (Table 4).
Improved ejection fraction Confidence Score (either from A to B, A to C, or B to C)
Harmonic imaging resulted in an improved ejection fraction Confidence Score in 13 of 70 patients (19%) when compared to standard imaging; contrast imaging resulted in an improved ejection fraction Confidence Score in 31 of 70 patients (44%) when compared to standard imaging (Table 4). This added value of contrast imaging over harmonic imaging (when each is compared to standard imaging) was statistically significant (p = 0.001). Contrast echocardiography also resulted in an improved ejection fraction Confidence Score compared to harmonic imaging in 26 of 70 patients (37%) (Table 4).
Different ejection fraction estimation (hyperkinesis, normal, three grades of hypokinesis)
Compared to standard imaging, harmonic imaging changed the estimation of ejection fraction in only 10 of 70 patients (14%); contrast imaging changed the estimation of ejection fraction as compared to standard imaging in 31 of 70 patients (44%) (Table 3). This added value of contrast imaging over harmonic imaging (each compared to standard imaging) was statistically significant (p < 0.0001). With contrast, the ejection fraction reading changed in both directions as compared to the other techniques (i.e., standard and harmonic imaging both overestimated and underestimated ejection fraction as compared with contrast imaging).
A subset of studies were randomly chosen to be re-read independently by two of the authors. A total of 128 left ventricular wall segments were assessed. On standard imaging there was agreement in 70% of wall segments evaluated. With harmonic imaging there was 80% agreement, and with contrast imaging there was agreement on 95% of the segments. The wall motion confidence scores agreed in 61% of cases on standard imaging, 70% on harmonic imaging, and 91% on contrast imaging.
Limitations of portable echocardiograms in the ICU setting
Echocardiography is an excellent noninvasive technique for the evaluation of left ventricular segmental and global systolic function. However, a certain percentage of transthoracic echocardiograms are technically suboptimal for this purpose. This is a special problem with regard to bedside, portable echocardiograms performed in the ICU. Patients in these locations are often on mechanical ventilation, which results in hyperinflated lungs that interfere with imaging. Even those not on ventilators often have pulmonary pathology. In addition, there often are objects that interfere with the placement of the transducer, such as bandages, chest tubes, and wires. Furthermore, ICU patients may not be able to cooperate with the sonographer by assuming an optimal position and holding their breath at end-expiration. The physical conditions in the ICU may be disadvantageous for obtaining echocardiographic images because of lighting and space considerations. Finally, the acuity of care in the ICU setting may prevent the sonographer from spending adequate time with the patient.
Overcoming the limitations of portable echocardiograms in the ICU setting
Recent innovations have improved the echocardiographic evaluation of left ventricular function. Among these are automatic edge-detection techniques such as Acoustic Quantification and Color Kinesis. However, both of these require good endocardial resolution, which frequently was not obtainable in our ICU patients. Harmonic imaging is a technique with which a higher frequency-matched complementary filter is used to allow reception of only the harmonic echoes inherent to tissue. Spurious echoes and artifacts may therefore be eliminated (14). Contrast imaging opacifies the cardiac blood pool, and thus allows endocardial border detection. This effect is enhanced when used in conjunction with harmonic imaging (15).
Transesophageal echocardiography overcomes many of the limitations of transthoracic echocardiography in evaluating wall motion. However, transesophageal echocardiography is more invasive, more expensive, uncomfortable, and has a small incidence of complications.
The current study has shown that the large majority of ICU patients (77%) were referred for echocardiography for the purpose of evaluating left ventricular systolic function. However, with standard, fundamental imaging, 34% of all left ventricular wall segments were uninterpretable with respect to wall motion in ICU patients. Left ventricular ejection fraction, perhaps the most sought-after single parameter of cardiac function, could not be obtained at all with standard echocardiographic imaging in 23% of our ICU patients, and could be obtained with surety in only 56%.
With the use of a safe and easily administered IV contrast agent, 76% of the left ventricular segments, which were uninterpretable with standard imaging, could be read with surety. Furthermore, 69% of the ICU patients who had uninterpretable ejection fractions with standard imaging had ejection fractions that could be read with surety on contrast imaging.
If this increase in confidence had not altered the actual results of wall motion and ejection fraction interpretation, the clinical impact of contrast imaging in this setting would be nil. However, contrast imaging actually changed the readings of both segmental and global wall motion in 44% of ICU patients, and also increased the readers’ confidence in their interpretations.
Interestingly, the added value of harmonic imaging without contrast was only marginal in this study of patients who were generally very difficult to image.
Although the readers were blinded to the clinical information and the results of their readings on standard and harmonic imaging when interpreting the contrast studies, no blinding occurred between the interpretations of the standard and the harmonic images. Therefore, no definite conclusion can be drawn about the paucity of improvement in interpretation and confidence with harmonic imaging over standard imaging.
As contrast imaging had previously been shown to be beneficial in other settings such as stress echocardiography, the readers could have been biased toward contrast imaging.
The improved later release of “tissue harmonic imaging” was not available during the time of this study.
The known limitations of contrast imaging (difficulty in visualizing the posterior wall on the parasternal long axis view and the lateral wall on the apical view because of shadowing) were encountered. Despite this, contrast did aid in visualization of the lateral wall, however, as this wall is especially hard to visualize on portable studies done in the ICU due to positioning and lung interference. Visualization of the anterior wall was also aided on contrast imaging.
Finally, there was no gold standard for the assessment of wall motion or ejection fraction in these patients.
Contrast imaging is a rapid, easily performed, and safe technique that dramatically changes both the confidence and results of segmental and global wall motion analysis in ICU patients. Therefore, this technique should be considered in all ICU patients whose transthoracic echocardiograms are not ideal for the evaluation of left ventricular function.
- intensive care unit
- Received December 7, 1998.
- Revision received July 22, 1999.
- Accepted October 18, 1999.
- American College of Cardiology
- Feinstein S.B,
- Cheirif J,
- Ten Cate F.J,
- et al.
- Porter T.R,
- Li S,
- Kilzer K,
- Deligonul U
- Porter T.R,
- Li S,
- Kricsfeld D,
- Armbruster R.W
- Villanueva F.S,
- Jankowski R.J,
- Manaugh C,
- Wagner W.R
- Bolognese L,
- Antoniuccu D,
- Rovai D,
- et al.