The pulmonary artery flotation catheter of Drs. Swan and Ganz is used primarily for monitoring hemodynamic parameters but also permits sampling of blood from the pulmonary circulation (Fig 1). With the exception of determinations of mixed venous oxygenation, little advantage has been taken of this unique access to blood from the pulmonary artery microcirculation.
Pulmonary microvascular disorders are notoriously difficult to diagnose in vivo. Amniotic fluid embolism and fat embolism classically have been definitively diagnosed only at autopsy. Lung biopsy is usually required for confirmation of the clinical suspicion of pulmonary microvascular and lymphangitic carcinomatosis. Histopathologic examination of the lung in patients with nonthrombotic pulmonary microembolism reveals abundant diagnostic material in terminal branches of the pulmonary artery circulation (Fig 2).We hypothesized that fetal squamous cells, fat globules, and malignant cells should be recognizable in blood extracted from the pulmonary circulation and, in particular, from microvascular blood obtained via the lumen of a pulmonary artery catheter in the “wedge” position (Fig 1).
Materials and Methods
Pulmonary artery and wedge positions are confirmed using conventional criteria. These include hemodynamic pressures and waveforms and analysis of blood gas levels. After discarding the first 3 ml of blood (the dead space of a 7 French thermodilution pulmonary artery catheter), 3 to 5 ml is collected anaerobically for analysis of blood gas levels and 10 ml for study.
Withdrawal of blood from the wedged catheter may be tedious. Gentle continuous aspiration is usually effective, but occasionally the balloon must be deflated and the catheter repositioned before an adequate volume of blood can be obtained.
In patients with suspected amniotic fluid embolism or lymphan-gitic carcinomatosis, samples are heparinized and centrifuged. Fetal squames and malignant cells are lighter than red blood cells and tend to accumulate in the bufly coat of centrifuged blood. Slides made from the bufly coat are immediately fixed with 95 percent alcohol and stained using the Papanicolaou method.
In our initial studies of patients with trauma, fat was found with equal ease after frozen sectioning and fat staining of plasma, whole blood, and blood clot; however, we have found that blood clot lends itself more readily to the frozen sectioning technique. Samples are allowed to clot, are decanted of serum, and then are frozen. Five-micron frozen sections are prepared with the frozen-sectioning microtome and are then processed with Sudan stain and evaluated microscopically for the presence of orange fat globules.
Care must be taken to clean the cryostat s blade with alcohol prior to sectioning of blood clot in order to remove any possible traces of fat from prior processing. We process control and study samples simultaneously to exclude the possibility of such artifacts.
Pulmonary Microvascular Findings Control Patients
We obtained samples of blood from ten patients who had no evidence of pulmonary diseases. In seven of these patients, pulmonary artery catheters had been inserted before surgery for hemodynamic monitoring. In three patients the indication for pulmonary artery catheterization was for guidance in fluid management following myocardial infarction.
Pulmonary arterial and wedge samples of blood were analyzed as described previously by frozen sectioning and fet staining and by Papanicolaou and Wright-Giemsa stains. Wright-Giemsa and Papanicolaou stains of smears of the bufly coat from pulmonary arterial blood showed normal elements of blood identical to those of peripheral venous blood; however, smears made from wedge samples invariably showed substantial numbers of megakaryocytes and megakaryocyte nuclei (Fig 3). These cells were frequently deformed and difficult to identify with certainty, particularly in Papanicolaou stained preparations.
Amniotic Fluid Embolism (Table 1)
We have found fetal squames in the pulmonary arterial blood of two young women with peripartum respiratory failure. In case 1, which has been previously reported, the adult respiratory distress syndrome (ARDS) developed after an uneventful vaginal delivery. In patient 2, toxemia and ARDS followed intrauterine fetal death. In both cases, pulmonary artery flotation catheters had been inserted by the staff of the intensive care unit for hemodynamic monitoring.
Cytologic analysis of samples of pulmonary arterial blood from both patients showed several components of amniotic fluid, including keratinized anuclear fetal squamous cells and mucous strands. The large diagnostic fetal squames were numerous and were easily distinguished from the background of smaller normal elements of blood (Fig 4). These cells are anuclear or contain pyknotic nuclei, as would be expected of mature epidermal cells.
Fat Embolism (Table 2)
We have studied seven patients who developed severe respiratory failure 12 to 72 hours after long-bone trauma. The clinical diagnosis of fat embolism had been suspected in patient 11, the one patient who fulfilled all three major clinical diagnostic criteria of the fat embolism syndrome, namely, respiratory insufficiency, diffuse neurologic dysfunction, and petechial rash. Other preliminary diagnoses in these patients included aspiration pneumonia, post-traumatic pulmonary insufficiency, and congestive heart failure. Chest roentgenograms in all but one patient showed a pattern of diffuse pulmonary edema, often with a predilection for the lower lung fields. Chest roentgenograms in patient 10 remained normal throughout her course. In this patient, samples of pulmonary arterial and pulmonary microvascular blood were taken at the time of pulmonary arteriographic studies performed because of suspected thromboembolic disease. In all other patients, pulmonary artery flotation catheters had been inserted to assist in the management of ARDS.Large numbers of fat globules were seen in samples of wedged blood from six patients. In patients 7, 8, and 11, fat was also seen in pulmonary arterial samples. The bright orange fat globules, varying in size from just a few micra to over 200|i, were easily detected among the background of hemolyzed frozen blood (Fig 5).
In patient 12, evidence of extensive external trauma to the chest wall and head suggested that contusions of the lungs and brain were responsible for respiratory and neurologic dysfunction, respectively. Fat studies in this patient were repeatedly negative. Alternative explanations for ARDS were not demonstrated in any other patients.
Lymphangitic Carcinomatosis (Table 1)
Two patients with biopsy-proven extensive lymphangitic and microvascular carcinomatosis (patients 3 and 5) gave informed consent for elective pulmonary artery catheterization to explore our hypothesis that pulmonary microvascular cytologic studies might yield malignant cells. In patient 4, samples were taken at the time of diagnostic right-sided heart catheterization to rule out pulmonary thromboembolism. This patient was thought to be too ill to tolerate fiberoptic bronchoscopy and transbronchial lung biopsy.
Pulmonary microvascular samples from patient 3 showed large numbers of malignant cells identical to those seen in biopsies of his lung and bone (Fig 6). Studies of samples of blood from patient 4 were very suspicious for cancer, but cellular distortion was such that not all cytologic criteria of malignancy could be met. In patient 5, clusters of large pyknotic cells were seen. We could not decide with certainty whether these represented cancer cells or megakaryocytic nuclei.
Figure 1. Pulmonary artery catheter in wedge position. PA, Pulmonary artery; and PCWP, pulmonary capillary wedge pressure.
Figure 2. Transbronchial lung biopsy, showing clusters of malignant cells (arrows) in pulmonary arteriole (case 5) (hematoxylin-eosin, original magnification X 400).
Figure 3. Pulmonary megakaryocyte (WWght-Giemsa stain, original magnification X900).
Figure 4. Pulmonary arterial blood from patient with amniotic fluid embolism. Several keratinized enucleate fetal squames are readily distinguishable from adjacent polymorphonuclear leukocytes (case 1) (Papanicolaou stain, original magnification x 400).
Figure 5. Fat globules from blood dot (Sudan stain, original magnification x400).
Figure 6. Malignant cells. Note thin rim of cytoplasm, two distinct nuclei, prominent nucleoli, and dense chromatin pattern (case 3) (Papanicolaou stain, original magnification x 400).
Table 1—Data on Amniotic Fluid Embolism and Lymphangitic Carcinomatosis
|Group and Case, Race, Sex, Age (yr)||Clinical Presentationt||Pulmonary Microvascular Cytologic Findings||Final Diagnosis||Disposition|
|Amniotic fluid embolism|
|1, W, F, 24||Postpartum ARDS||4 + fetal squames||AFE||Alive and well;|
|2, W, F, 18||ARDS secondary to toxemia and retained dead fetus||4 + fetal squames||AFE||no residua|
|3, W, M, 66||Dyspnea; normal CXR;? pulmonary embolism; ? LC||4 + class 5, adenocarcinoma||LC (transbronchial lung biopsy)||Sustained remission (1 yr); then death from pulmonary recurrence|
|4, W, M, 64||Metastatic renal cell cancer to mediastinum (open biopsy); increased dyspnea; bilateral densities on CXR; LC vs pulmonary embolism vs radiation pneumonitis||3+ class 4, probable malignant cells||Probable LC||Normal pulmonary arteriogram; rapid death|
|5, W, M, 58||U ndifferentiated||Large pyknotic cells, not||LC (transbronchial||Noninvasive supportive|
|bronchogenic carcinoma of RUL, with LC||classifiable||lung biopsy)||therapy; rapid death|
Table 2—Data on Fat Embolism
|Case, Race, Sex, Age (yr)||Central||Sudan Stain*||FinalDiagnosis|
|6, W, M, 38||Pelvis; tibia; fibula; mandible||4 +||2 +||No||0||2 +||Fat embolism|
|7, W, F, 72||Sternum; tibia||2 +||2 +||No||2 +||4 +||Fat embolism|
|8, W, F, 70||Feet; pelvis; spine||1 +||4 +||No||2 +||4 +||Fat embolism|
|9, W, F, 69||Femur; ribs||0||3 +||No||0||4 +||Fat embolism|
|10, W, F, 70||Hip||2 +||3+ (decreased Po2; clear CXR)t||No||0||3 +||Fat embolism; COPD|
|11, W, F, 84||Hip||2 +||4 +||3 +||3 +||3 +||Fat embolism|
|12, W, M, 21||Pelvis; spine||2 +||3 +||No||0||0||Pulmonary and cerebral contusions|