Reduction in Left Ventricular Filling During Bronchopulmonary Lavage Demonstrated by Transesophageal Echocardiography

Abstract

Pulmonary alveolar proteinosis (PAP) is an uncommon disease first described in 1958 as an accumulation of lipoproteinacious material within the alveoli [1]. The etiology of PAP is uncertain and various forms, including idiopathic and disease secondary to dust or fume exposure, have been reported [1-3]. Although it is classically described as involving the alveoli, PAP can also be associated with septal edema and interstitial fibrosis [3,4]. In 1963, Ramirez [5] reported bronchopulmonary lavage (BPL) as a therapy for PAP and, since that time, it has remained the treatment of choice. Most patients have marked improvement in symptoms after a single treatment, but require repeated lavage over months or years to maintain a therapeutic effect [6,7]. BPL is a unique procedure which requires general anesthesia and separation of the lungs with a double-lumen tube. One lung is then filled with up to 1500 mL of normal saline, which is infused by gravity from a position 30 cm above the patient's midaxillary line. The saline is then drained from the lung by gravity into a container which is 60 cm below the midaxillary line [5]. This process is usually repeated until the effluent is clear. The infusion of large volumes of saline into the lung can be associated with significant hemodynamic changes. Hypotension [5], increased central venous pressure [7] and changes in arterial oxygenation [8] have all been reported to occur during the infusion of large volumes of saline into one lung. Invasive hemodynamic monitoring performed during BPL has demonstrated a consistent pattern of increased central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP) while the lung is filled with saline [7,8]. Despite these increases in CVP and PCWP, it has been postulated that ventricular filling is not increased, and may actually be compromised when large volumes of saline are infused into the lung [8,9]. The determination of ventricular preload with pulmonary artery or central venous catheters can be misleading since pressure measurements do not consistently reflect ventricular volume and may be affected by other factors such as intrathoracic pressure [10,11]. In order to record accurate trends in ventricular preload during BPL, we measured the change in the left ventricular short axis area at the midpapillary level during BPL using transesophageal echocardiography (TEE). Short axis measurements of the inferior vena cava (IVC) at a level immediately superior to the hepatic veins were also obtained to record the effect of BPL on central venous filling. Automated border detection (ABD), a technique capable of providing real time estimation of cavity areas [12], was used to provide objective, serial measurements of left ventricular and inferior vena caval dimensions. Case Report A 35-yr-old, 50-kg, 160-cm male with PAP was scheduled for BPL of the left lung. The patient had been treated with successive right and left lung lavage 1 mo earlier resulting in significant improvement of his pulmonary function. A review of his anesthetic records during the prior lavages was remarkable for periods of hypotension while the lung was filled which resolved as the lung was emptied. Systolic blood pressure had decreased from a range of 110-120 mm Hg, with the lung empty, to 60-65 mm Hg with the lung full. Because of this prior evidence of hemodynamic instability during BPL, the TEE was used during the procedure to monitor cardiac function. After anesthetic induction with alfentanil 50 micro gram/kg and propofol 1.0 mg/kg, the patient received rocuronium 0.4 mg/kg for neuromuscular block and a 37 mm, left-sided, double-lumen endotracheal tube was placed without difficulty. Correct position was verified using auscultation and direct visualization with fiberoptic bronchoscopy. A 20-gauge cannula was placed in the left radial artery to provide continuous monitoring of arterial blood pressure and permit intermittent blood gas tension analysis. Anesthesia was maintained using a continuous intravenous infusion of alfentanil 0.75-1.5 micro gram centered dot kg-1 centered dot min-1 and propofol 60-120 micro gram centered dot kg-1 centered dot min-1 as required. The left lung was deflated and the patient was mechanically ventilated with 100% oxygen. Ventilatory variables during one lung ventilation included a tidal volume of 500 mL at a rate of 12 breaths/min. The peak inspiratory pressure varied between 33-38 mm Hg. The patient remained supine during the entire procedure. Prior to beginning BPL, a two-dimensional omniplane TEE probe (Model 21364A; Hewlett Packard, Andover, MA) was placed without difficulty. After recording a basic examination, the probe was positioned at the transgastric, short axis, midpapillary view of the left ventricle and the highest quality conventional image was obtained. The real-time ABD system was then activated and the operator adjusted the time-gain compensation controls to superimpose the automated borders with the visually apparent endocardial borders. Measurements of the left ventricular end-diastolic area (EDA) were obtained with the lung full and empty during three consecutive lavage cycles. All areas were recorded at end-expiration with mechanical ventilation off for 15 s. Simultaneous measurements of the arterial blood pressure were recorded. After recording the left ventricular areas, the probe was withdrawn slightly and rotated in a clockwise direction until a short axis image of the IVC was obtained at a level immediately superior to the hepatic veins. ABD was again used to record the short axis area during two lavage cycles. The results of these measurements are demonstrated in Table 1. Figure 1 and Figure 2 demonstrate the left ventricular and vena caval areas, respectively, as recorded using ABD with the lung full and empty.Table 1: Left Ventricular and Inferior Vena Caval Echocardiographic Measurements During Lung LavageFigure 1: A and B. The end-diastolic area (EDA) of the left ventricle with the lung empty (A) and saline filled (B) during lavage cycle 1. A 49.2% reduction in EDA was calculated using automated border detection (ABD).Figure 2: A and B. The inferior vena cava (IVC) at the level of the hepatic veins with the lung empty (A) and saline filled (B) during lavage cycle 5. Using automated border detection, a 140% increase in IVC short axis area was noted as the lung was filled with saline.The patient's systolic blood pressure during five intervals with one lung deflated ranged from 88 to 110 mm Hg. During five corresponding intervals while one lung was fluid filled, the range of systolic blood pressure was 52-67 mm Hg. On two occasions, approximately 20 min prior to ABD measurements, the patient received ephedrine 10 mg intravenously to support blood pressure while the lung was fluid filled. The patient's oxygen saturation was noted to increase from a range of 89%-92%, during periods when the lung was empty, to 95%-97% when the lung was saline filled. After completing 15 total liters of lavage in 1000-mL increments, the effluent was judged to be sufficiently clear and the left lung was re-expanded and ventilated. After anesthetic emergence, the patient's trachea was extubated without incident. Discussion Hemodynamic changes associated with BPL were first reported in the original description of this procedure by Ramirez [5], who noted significant hypotension in a patient each time the left lung was filled with saline. In 1970, Smith et al. [8] reported significant increases in both CVP and arterial oxygen tension during periods when the lavaged lung was filled with saline. These same authors also measured increased pulmonary arterial, esophageal, and left ventricular end-diastolic pressures in an animal model during periods when the lung was fluid-filled. They postulated that these changes were caused by a generalized increase in intrathoracic pressure while the lung was full. The increase in arterial oxygen tension was believed to result from a decrease in pulmonary capillary perfusion and shunt in the fluid-filled lung, with a concomitant increase in blood flow to the ventilated lung. In 1977, Busque [7] also reported increased CVP which he attributed to an increase in intrathoracic pressure and a shift of the mediastinum away from the fluid-filled lung. Increased CVP and PCWP do not explain the presence of hypotension in PAP patients with normal left ventricular function during BPL. We therefore elected to directly measure left ventricular EDA as well as extrathoracic inferior vena caval short axis areas to provide an accurate measure of left ventricular filling and function. The ABD system was used to provide objective, real-time measurement of end-systolic and EDA without the need to manually trace the endocardial border during each lavage cycle. ABD has been validated as a method of providing accurate, real-time measurements of cavity areas [12-14]. Because the results are based on fixed gain settings, endocardial borders are consistent and not subject to variation seen with serial manual tracings. The results in this case show a reduction in left ventricular EDA Figure 1b of up to 61% Table 1, suggesting that the ventricle may be significantly underloaded when the lung is fluid filled. Examination of the right heart did not reveal septal flattening, ventricular dilatation, tricuspid regurgitation, or other evidence of acute pressure overload or right ventricular dysfunction. The decrease in ventricular preload in our patient seems to be related, at least in part, to a decrease in venous return to the heart. The IVC, measured in short axis below the diaphragm, was markedly dilated during periods when the lung was full Figure 2b. Measurements of the IVC using ABD demonstrated an increase in area of up to 140% during lung filling Table 1. These data would suggest impaired venous return to the thorax of uncertain etiology. Increased intrathoracic pressure and mediastinal shift have been suggested by other authors as an explanation for the hemodynamic changes associated with lung lavage [7-9]. The left ventricular EDA reported in this patient with the lung deflated Table 1 was 5.9 to 7.1 cm2 and is lower than mean EDA areas reported in other series using ABD [12,13,15]. The lower left ventricular EDA measured in our report is within the range of normal and may be expected considering the patient's less than average height and weight. In summary, using transesophageal echocardiography, we were able to identify and document the association of decreased left ventricular EDA with recurrent hypotension in a patient with PAP undergoing BPL. In the past, pulmonary artery occlusion and central venous pressure measurements have been misleading with regard to changes in left ventricular preload as a contributing cause of hypotension. This case also illustrates the utility of ABD as a valuable tool for continuous quantification of changes in ventricular preload.