Hospital Practice: Volume 37: No.1

Sumit Verma, MD; Sumit Kumar, MD; James R. Gossage, MD; And Vipul B. Shah, MD

Abstract: A prospective study was performed on the utility of echocardiography in diagnosing hypotension in critically ill patients. In our study, we found that transthoracic echocardiography can help physicians determine the etiology of hypotension in a significant number of patients. Transesophageal echocardiography is useful when results obtained from transthoracic echocardiography are suboptimal. Left ventricular function assessed by echocardiography can be used to predict 30-day mortality.

Keywords: hypotension; echocardiography; transthoracic echocardiography; transesophageal echocardiography

Hypotension frequently develops in critically ill patients. Common causes of hypotension in patients in the intensive care unit (ICU) include left ventricular dysfunction due to myocardial infarction, hypovolemia, sepsis, and pulmonary embolism. Echocardiography is a diagnostic test that is easy to perform and is useful in the early diagnosis of the foregoing abnormalities. Previous studies that used echocardiography in patients with hypotension had limitations. One of the limitations is a selection bias that occurs because echocardiography is generally used for evaluation of hypotension in patients with known or suspected cardiac abnormalities only, so an echocardiogram may not be done on a “noncardiac” patient. It is certainly possible that other subgroups of patients may benefit from echocardiographic evaluation.¹ We prospectively studied the diagnostic utility of transthoracic echocardiography (TTE) in unselected consecutive patients with persistent hypotension who were admitted to the ICU. Patients were enrolled in 4 ICUs: 2 medical and 2 cardiac ICUs. We also studied the incremental value of transesophageal echocardiography (TEE) in patients who had inadequate results from TTE studies.

All patients with persistent hypotension admitted to the ICU were eligible for inclusion. We prospectively enrolled patients from 4 separate ICUs in 2 hospitals: 2 medical and 2 coronary. There were 80 men and 37 were women. In 2 ICUs located in a Veterans Administration hospital, no women were enrolled. The mean age was 61 years (range, 21–91 years).

Persistent hypotension was defined as the persistence of either of the following situations for at least 2 hours: 1) a fall in systolic arterial pressure to < 90 mm Hg; 2) need for vasopressor agents such as dopamine 5 μg/kg/min or any other vasopressor agent to maintain a systolic arterial pressure of ≥ 90 mm Hg.

All patients were enrolled in accordance with local Institutional Review Board regulations. Informed consent for enrollment in the study was not required. Informed consent for TEE was obtained. Patients were identified during daily rounds in the ICU by members of the research team who reviewed the charts of all admitted patients. If a patient was admitted after morning rounds, he or she was identified by the clinical team managing that patient. In addition, the ICU patient log book was checked daily to identify patients who were admitted during the evening hours. All vital signs were recorded from ICU flow sheets. Transthoracic echocardiography was performed within the first 24 hours after the diagnosis of hypotension. The echocardiogram was interpreted by an echocardiographer who was not involved in clinical decision making. Transesophageal echocardiography was performed when the results of the TTE study were inadequate, or if the team deemed it necessary for any other reason for clinical management of the patient.

The following criteria were defined prospectively as necessary to establish a definitive diagnosis for the etiology of hypotension using clinical criteria. The clinical team was not blinded to the echocardiographic findings. Multiple diagnoses were allowed if they were made by the clinical team. These diagnoses were compared with those suggested by echocardiography.

Septic shock was diagnosed if both of the following criteria were met: 1) ≥ 3 systemic inflammatory response syndrome (SIRS) criteria: heart rate ≥ 90 beats per minute, core temperature ≥ 38°F or ≤ 36°F, respiratory rate ≥ 20/min, white blood cell count ≥ 12 000/mm or ≤ 4000/mm or 10% band forms;² 2) a suspected clinical site of infection based on objective clinical data (eg, infiltrates evident on a chest radiograph, pyuria, positive results from a blood culture, or an abscess on imaging).

Hypovolemia was diagnosed if one of the following criteria was met: 1) clinical scenario consistent with hypovolemia—presence of gastrointestinal bleeding or history compatible with volume depletion (for example, diuresis, dehydration, retroperitoneal hemorrhage); 2) central venous pressure < 10 mm Hg, pulmonary artery wedge pressure < 12 mm Hg, and associated cardiac index < 2.5/min/m²; 3) resolution of hypotension with fluid resuscitation.

A diagnosis of pneumothorax required the presence of both of the following criteria: 1) pneumothorax visible on a chest radiograph; 2) resolution of hypotension after chest tube insertion.

Auto-positive end-expiratory pressure (auto-PEEP) or ventilator-induced hypotension required both of the following criteria to be met: 1) auto-PEEP detected on ventilator or hypotension after initiation of mechanical ventilation; 2) resolution of hypotension with correction of auto-PEEP.

A patient was diagnosed with a pulmonary embolism if any one of the following criteria was met: 1) high-probability ventilation-perfusion (V/Q) scan with > 30% defects; 2) positive results on a computed tomographic scan for central pulmonary embolus; 3) positive results on an angiogram with > 30% pulmonary vascular obstruction; 4) intermediate probability ventilation perfusion V/Q scan with > 30% defects and positive results from a Doppler examination of the lower extremity; 5) unequivocal clot visible on TEE with right ventricular volume overload and persistent hemodynamic instability.

Drug-induced hypotension was diagnosed when both of the following criteria were present: 1) documented (by history or medication records) receipt within 3 hours preceding onset of hypotension of a drug commonly associated with hypotension; 2) resolution of hypotension with fluid resuscitation, drug withdrawal, or administration of antagonist where available.

Acute myocardial infarction was diagnosed when both of the following criteria were present: 1) chest pain or isch-emic electrocardiographic changes (ST elevation ≥ 1 mm in 2 consecutive chest leads or ST elevation ≥ 2 mm in limb leads), new Q waves, or left bundle branch block; 2) acute rise and fall in the levels of creatine phosphokinase and in the creatine kinase-MB fraction or a positive result from the troponin T assay.

The diagnosis was cardiogenic shock if all of the following criteria were met: 1) left ventricular ejection fraction (LVEF) ≤ 25% and bilateral pulmonary infil-trates on a chest radiograph; 2) pulmonary artery wedge pressure ≥ 18 mm Hg and cardiac index < 2.5 L/min/m²; 3) echocardiographic evidence of severe valvular dysfunction and bilateral pulmonary infiltrate visible on a chest radiograph.

Pericardial tamponade was diagnosed if pericardial fluid was visible on an echocardiogram and any one of the following criteria was present: 1) physiologic signs of tamponade visible on an echocardiogram; 2) pulmonary artery catheterization hemodynamics consistent with tamponade; and 3) resolution of hypotension with removal of pericardial fluid.

Arrhythmia was diagnosed if either 1 or 2 was present together with 3: 1) ventricular tachycardia, ventricular fibrillation, or bradycardia with a heart rate ≤ 50 bpm; 2) atrial fibrillation or supraventricular tachycardia with a heart rate of > 150 bpm; and 3) resolution of hypotension with correction of the arrhythmia.

Multiorgan failure was defined for kidney, liver, lung, and disseminated intravascular coagulation according to criteria established by the American College of Chest Physicians and the Society of Critical Care Medicine.³

Complete TTE (2-dimensional, M-mode, Doppler, and color flow Doppler) was performed to evaluate the etiology of the hypotension. If TTE was technically inadequate or the clinical team felt it was necessary, TEE was performed if none of the usual contraindications were present. The results were reviewed by an independent reviewer, and only the echocardiographic findings were used to determine a probable reason for the patient’s hypotension. Cardiac tamponade, severe ventricular dysfunction, and severe valvular insufficiency of the mitral and aortic valves were considered significant findings that explained the etiology of the hypotension. Severe left ventricular dysfunction (LVEF < 30%) was classified as preexisting or new cardiomyopathy, depending on whether the echocardiographic findings existed before hospitalization. Patients who had wall motion abnormality in an area served by a single coronary artery were classified as having “possible myocardial infarction” (realizing that echocardiographic findings do not indicate date of infarction). Patients with a small left ventricular cavity with hyperdynamic left ventricular function were considered to have hypovolemia. Patients with right ventricular enlargement and elevated pulmonary artery systolic pressure were classified as having right ventricular overload. Other possible etiologies were determined on an individual basis.

The quality of the echocardiogram was classified as adequate if data could be obtained from the following processes: 1) evaluation of left and right ventricular function and assessment of wall motion in all segments; 2) assessment for pericardial effusion; 3) evaluation of valvular function using color flow Doppler echocardiography.

The echocardiogram was classified as partially useful if the image quality was suboptimal but some abnormality was identified that indicated the etiology of the hypotension. The echocardiogram was classified as inadequate if no useful information was obtainable because of poor image quality. All patients enrolled were assessed for outcome mortality at 30 days.

We prospectively identified 144 consecutive patients with hypotension, 117 of whom had persistent hypotension. Significant abnormalities indicating the etiology of hypotension were identified by echocardiography in 93 of 117 (79.4%) patients; the etiologies of the remaining 24 (20.6%) patients with hypotension were not identifiable by echocardiography. Of these 117 patients, 56 had a normal or a hyperdynamic LVEF (≥ 55%); 40 had an LVEF between 26% and 54%; the remaining 21 had an LVEF ≤ 25% (Figure 1). The abnormalities identified by echocardiography were classified as defined previously. The distribution of the abnormalities is shown in Figure 2.

Hypovolemia was diagnosed in 6 patients by the presence of a small hyperdynamic left ventricular cavity. Fourteen patients were diagnosed as having new cardiomyopathy, and 15 patients had echocardiographic findings indicating preexisting cardiomyopathy. Possible myocardial infarction was suggested in 28 patients; 24 of the 28 patients had acute myocardial infarction diagnosed by subsequent clinical correlation with electrocardiographic and enzymatic abnormalities. The remaining 4 patients were diagnosed as having had a remote myocardial infarction on the basis of the clinical evaluation. Other patients had severe aortic regurgitation (n = 4), severe mitral regurgitation (n = 3), tamponade (n = 4), and right ventricular overload due to pulmonary embolus (n = 6).

Echocardiography disclosed significant information that was not otherwise suspected in several patients, and these results altered patient management.

1. In a patient with suspected urosepsis, findings of severe left ventricular dysfunction led to specific therapy directed at ventricular dysfunction with the addition of inotropic agents.

2. In a patient with hypotension (thought to be caused by intravenous nitroglycerin), an echocardiographic finding of severe aortic regurgitation led to aortic valve replacement surgery.

3. In a patient who had undergone percutaneous coronary intervention, echocardiography indicated previously unsuspected severe aortic regurgitation that was found to be due to infective endocarditis.

4. In 3 patients with suspected sepsis, left ventricular outflow tract obstruction related to hyperdynamic left ventricular function led to discontinuation of vasopres-sors. This finding also allowed for hypotension to be treated with intravenous fluids and a β-blocker, with recovery from hypotension. This unusual etiology of hypotension can occur in response to inotropic therapy. The correct diagnosis of this condition is not possible without echocardiography.

5. Transesophageal echocardiography led to the discovery of an unusual case of cardiac tamponade caused by right arterial compression by pleural fluid that caused hypotension.

Echocardiographic findings were consistent with the final clinical diagnosis in 105 of 117 (89.7%) patients. It should be noted, however, that the clinical team was not blinded to the echocardiographic findings. We did not record how frequently echocardiographic results led to alterations in medications or other invasive procedures. Patients with normal left ventricular function had numerous diagnoses (Table 1). The mean LVEF in patients with abnormal echocardiographic results was 32.6%.

View: (Table 1 ) - Clinical Diagnoses of Etiology of Hypotension in Patients with Normal Left Ventricular Function
Transesophageal Echocardiography

Transesophageal echocardiography was performed in 12 patients following inadequate TTE (Table 2). In one patient, TEE was recommended but not performed because the patient had a tracheoesophageal fistula.

View: (Table 2 ) - Transesophageal Echocardiographic Findings in Patients with Hypotension
Statistical Analysis

Analysis of data using LVEF in relation to the 30-day mortality outcome was performed. Of 117 patients, 75 (group 1) survived 30 days; the remaining 42 (group 2) did not survive. We performed a 2-sample t-test comparing the mean LVEF of group 1 with that of group 2. The mean LVEF of patients who survived was significantly higher than the mean LVEF of nonsurvivors (P = 0.005), with a t value of 2.62.

We then calculated a cutoff point for the LVEF that may help predict a patient’s chances of survival. We derived the following linear discriminate function:

g ( X ) = 0.2094 × X ­ 5.4649 , in which X = LVEF value .

The classification procedure is as follows: classify a future patient in the first group if g (X) > 0 (ie, if the LVEF is > 26%). Otherwise, classify a patient in the second group. This classification provided a total correction classification rate of 70.1%.

We also noted that when the LVEF was ≤ 25%, 14 of 21 patients did not survive (mortality rate, 66.7%) (Figure 3). When the LVEF was > 25% and < 55%, 12 of 38 patients did not survive (mortality rate, 31.6%). When the LVEF was ≥ 55%, 16 of 58 patients did not survive (mortality rate, 27.5%). In the last 2 categories, the difference was not statistically significant. Therefore, these categories were combined. Thus, when the LVEF was > 25%, 28 of 96 patients did not survive (mortality rate, 29.2%). To test for a significant difference in mortality rate between patients with hypotension with a LVEF ≤ 25% compared with a LVEF ≥ 25%, we performed a z test:

View: (Figure 3 ) - Thirty-day mortality rate in patients, depending on their LVEF.
z score = 3.32364 , with a P value of 0.00044 < 0.05.

We concluded that there was a significant difference in mortality rate in patients with hypotension and with an LVEF ≤ 25% compared with an LVEF ≥ 25%. This result correlates with the results obtained using linear discriminant analysis.

We also performed a logistic regression analysis to obtain the probability of classifying a future patient in a particular group. Our derived probability of classifying a new patient in group 1 using the logistic regression model is given by P (classifying a new patient into G1) = EXP (−0.96 + 0.035 × X)/1 + EXP(−0.96 + 0.035 × X), in which X = LVEF. The area under the ROC curve is 70.1%.

Using the preceding logistic regression model, we also concluded that if the LVEF is ≤ 25%, the chance of surviving another 30 days is < 47.9%; if the LVEF is > 26% and < 54%, the survival rate will be between 48.8% and 71.7%. If the LVEF is ≥ 55%, then the survival rate will be > 72.4%.

The 30-day mortality rate was high in these patients, regardless of ventricular function. Patients who did not survive 30 days had a mean survival of 14 days (range, 1–29 days). A significant number of patients (49.6%; n = 58) in this series had left ventricular systolic dysfunction.

Discussion

Echocardiography is a useful diagnostic method for evaluating hypotension.^4-6 It permits the diagnosis of a variety of cardiac and noncardiac abnormalities that result in hypotension. It can help diagnose conditions such as hypovolemia and pulmonary embolism.^7,8 The algorithm we used is easy to apply clinically.

Transesophageal echocardiography has advantages over TTE because it provides better images of structures such as the aorta that are not well visualized even with high-quality TTE studies.^9,10 Previous studies have shown that TEE may be superior to TTE.¹ However, TEE is difficult to perform on all patients for several reasons. In critically ill patients, the risk of problems with passing the TEE probe is increased, and sedating patients who are awake is risky because the drugs can worsen hypotension. In the study by Heidenreich et al,¹ 3 of 61 patients suffered a minor complication from TEE. In our series, we had a high percentage of adequate transthoracic studies. We attribute this to experienced technicians and to the definition we used for adequate. We did not require that all echocardiographic views be adequate. An adequate study was one in which we were able to collect all of the data required. Using this definition, we classified 83% of the TTE studies as adequate, 12% as partially useful, and 5% as inadequate. In patients in whom TEE was performed, useful information was obtained in most cases. The data suggest that a diagnosis can be obtained by using TTE in a large number of patients with hypotension. We found that 79.4% of patients had abnormal results on an echocardiogram that suggested a diagnosis of hypotension. Many patients had diagnoses not suspected from clinical findings. We identified left ventricular outflow tract obstruction in 6 of 117 patients that resulted from the use of inotropic agents.¹¹ This finding was perhaps the most surprising in this study because it identified an etiology of hypotension that required discontinuation rather than higher doses of vasopressors. Without the results from the echocardiographic scans, it is possible that higher doses of vasopressors could have been administered, leading to worsening of hypotension. No particular patient characteristics such as age or gender predicted this finding. In all cases, improvement occurred with discontinuation of the vasopressor agent and infusion of β-blockers and intravenous fluids. This particular etiology of hypotension can only be diagnosed with echocardiography, which emphasizes the importance of this modality for evaluating patients with hypotension.

Echocardiography can be useful in diagnosing hypovolemia.^7,8 We found a significant number of cases in which echocardiography provided hemodynamic information that allowed appropriate fluid management. It is possible that echocardiographic results confirmed clinical suspicions in several patients, which may have prevented some unnecessary interventions or diagnostic tests. For ethical reasons, we did not blind the management team to the echocardiographic results, hence the high degree of concordance of clinical diagnosis with the echocardiographic findings. The significance of left ventricular dysfunction in the eventual outcome is emphasized in this study. We found that a significantly worse outcome (30-day mortality) is likely if the LVEF is ≤ 25% than if it is > 25%.

Limitations

In comparison with previous studies, this study included consecutive patients with hypotension; hence, it avoided selection bias. All patients who qualified based on entry criteria were enrolled and followed. The number of TEEs performed (n = 12) was relatively low because we conducted a large number of TTE studies. Transesophageal echocardiographic studies were only performed as an adjunct when TTE was inadequate. It is possible that additional information would have become available if all patients had undergone TEE scans. Our findings help identify the incremental value of TEE over TTE. It has been suggested that TTE should be eliminated from the diagnostic algorithm and that TEE should be performed directly. Our findings suggest otherwise. Also, if a TTE scan is adequate, it is not necessary to subject a patient to the potential risk of TEE. Transthoracic echocardio-graphic studies were performed by experienced technicians. Unfortunately, we did not record which patients were mechanically ventilated at the time of echocardiography, which is occasionally the reason for an inadequate study.

The clinical team was not blinded to the results of echocardiography. In fact, findings were made available to them to help in clinical decision making. This fact may explain the high degree of concordance between clinical and echocardiographic diagnoses. For this reason, the relationship between a diagnosis based on echocardiography alone and the final clinical diagnosis is difficult to ascertain. This study provides us with the potential value of echocardiography in an unselected population. However, the current cost-conscious environment requires careful clinical selection of this and other tests. Echocardiography does not by any means replace good clinical management, but it is likely to enhance management of hypotensive patients.

According to the American Society of Echocardiography’s Appropriateness Criteria for Transthoracic and Transesophageal Echocardiography,¹² an appropriate imaging study is one in which the expected incremental information, combined with clinical judgment, exceeds the expected negative consequences by a sufficiently wide margin for a specific indication that the procedure is generally considered acceptable care and a reasonable approach for the indication. They assign a score between 1 and 9 for appropriateness for indications for TTE and TEE. Clinical evaluation of hypotension is assigned a score of 9 (appropriate). However, TTE/TEE for initial evaluation of pulmonary embolism is assigned a score of 3 (inappropriate).

Conclusion

This study assessed the utility of echocardiography in determining the cause of hypotension in critically ill patients. In a population of unselected patients, TTE provided useful diagnostic information rapidly in a large percentage of patients.

Echocardiography can identify causes of hypotension that are not clinically suspected and can alter clinical management. In addition, determining left ventricular function at the onset of hypotension helps predict patient outcome in respect to 30-day mortality. We were not able to determine how many deaths, if any, may have been prevented by using echocardiography.

Acknowledgments
The authors would like to acknowledge the contribution of Subhash C. Bagui, PhD in providing assistance in statistical analysis.
Hello!

Conflict of Interest Statement
Sumit Verma, MD, Sumit Kumar, MD, James R. Gossage, MD, and Vipul B. Shah, MD disclose no conflicts of interest.
References
1. Heidenreich PA, Stainback RF, Redberg RF, Schiller NB, Cohen NH, Foster E. Transesophageal echocardiography predicts mortality in critically ill patients with unexplained hypotension. J Am Coll Cardiol. 1995;26(1):152–158.


2. Rangel-Frausto MS, Pittet D, Costigan M, Hwang T, Davis CS, Wenzel RP. The natural history of systemic inflammatory response syndrome (SIRS). A prospective study. JAMA. 1995;273(2):117–123.


3. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ ATS/SIS International Sepsis Definitions Conference. Crit Care Med. 2003;31(4):1250–1256.


4. Donovan KD, Colreavy F. Echocardiography in intensive care: the basics. Part I. Crit Care Resusc. 1999;1(3):291–295.


5. Donovan KD, Colreavy F. Echocardiography in intensive care: the basics. Part II. Crit Care Resusc. 1999;1(3):296–310.


6. Stanko LK, Jacobsohn E, Tam JW, De Wet CJ, Avidan M. Transthoracic echocardiography: impact on diagnosis and management in tertiary care intensive care units. Anaesth Intensive Care. 2005;33(4):492–496.


7. Jensen MB, Sloth E, Larsen KM, Schmidt MB. Transthoracic echo-cardiography for cardiopulmonary monitoring in intensive care. Eur J Anaesthesiol. 2004;21(9):700–707.


8. Vignon P. Hemodynamic assessment of critically ill patients using echocardiography Doppler. Curr Opin Crit Care. 2005;11(3):227–234.


9. Colreavy FB, Donovan K, Lee KY, Weekes J. Transesophageal echocardiography in critically ill patients. Crit Care Med. 2002;30(5): 989–996.


10. Khoury AF, Afridi I, Quiñones MA, Zoghbi WA. Transesophageal echocardiography in critically ill patients: feasibility, safety, and impact on management. Am Heart J. 1994;127(5):1363–1371.


11. Brown JM, Murtha W, Fraser J, Khoury V. Dynamic left ventricular outflow tract obstruction in critically ill patients. Crit Care Resusc. 2002;4(3):170–172.


12. ACCF/ASE/ACEP/ASNC/SCAI/SCCT/SCMR 2007 appropriateness criteria for transthoracic and transesophageal echocardiography: a report of the American College of Cardiology Foundation Quality Strategic Directions Committee Appropriateness Criteria Working Group, American Society of Echocardiography, American College of Emergency Physicians, American Society of Nuclear Cardiology, Society for Cardiovascular Angiography and Interventions, Society of Cardiovascular Computed Tomography, and the Society for Cardiovascular Magnetic Resonance. Endorsed by the American College of Chest Physicians and the Society of Critical Care Medicine, et al. J Am Soc Echocardiogr. 2007;20(7):787–805.


Sumit Verma, MD ¹
Sumit Kumar, MD ²
James R. Gossage, MD ³
Vipul B. Shah, MD <sup>4

1</sup>Regional Heart and Vascular Institute, Pensacola, FL ²Drexel University College of Medicine, Philadelphia, PA ³Medical College of Georgia, Augusta, GA ⁴Carolina Cardiology Associates, Rock Hill, SC

Correspondence: Sumit Verma, MD, Regional Heart and Vascular Institute, Pensacola, FL 32504.,
Tel: 850-382-1476,
E-mail: verma.ep@gmail.com
Disclaimer
In an effort to provide information that is scientifically accurate and consistent with accepted standards of medical practice, the editors and publisher of Hospital Practice routinely consult sources believed to be reliable. However, readers are encouraged to confirm this information with other sources. For example and in particular, physicians are advised to consult the prescribing information in the manufacturer's package insert before prescribing any drug mentioned.





Back to the table of contents for the December 2009 issue