Case of Exogenous Lipoid Pneumonia: Steroid Therapy and Lung Lavage with an Emulsifier

A 44-yr-old woman with schizophrenia (medical history otherwise negative) was admitted to the intensive care unit with acute respiratory distress syndrome. Mechanical ventilation was started. Positive end-expiratory pressure, 15 cm H²O, resulted in a marked improvement in oxygen fraction ratio from 107 to 369 (Fio²= 0.7), Paco²from 45.9 mmHg to 41.6 mmHg, and respiratory compliance from 46.7 to 60.9 ml/cm H²O. The patient was initially given wide-spectrum antibiotics, which were discontinued after 4 days because microbiologic cultures of tracheal aspirate, bronchoalveolar lavage fluid, blood, and urine were negative and remained negative throughout the clinical course. Vaseline oil intoxication was diagnosed on day 2, and methylprednisolone was given (2 mg · kg⁻¹· day⁻¹).4The severe respiratory failure steadily improved for 20 days, but when we discontinued methylprednisolone on day 29 gas exchange deteriorated, and the patient became hemodynamically unstable and presented septic shock without infection. Methylprednisolone was restarted (2 mg · kg⁻¹· day⁻¹, day 30), with improvement of hemodynamics and respiratory function. However, a subsequent attempt to taper the methylprednisolone to 0.25 mg · kg⁻¹· day⁻¹on day 44 again resulted in deterioration. The quantitative analysis results of a lung computed tomographic (CT) scan taken on day 46 were nearly identical to those of the scan taken on day 1 (fig. 1). We measured the vaseline oil concentration in the lung secretions by nuclear magnetic resonance and infrared spectrometry, and we found a concentration of 44 mg/ml (no data are available in the literature for comparison). Because the steroid therapy had failed to cure the disease and because we felt that it was unlikely that isotonic saline would remove the immiscible oil, we looked for an agent that could be added to the lavage solution to facilitate removal of the hydrocarbons.

After several in vitro  tests (see Discussion), we concluded that the best agent to emulsify the vaseline oil secretions in the lung was a solution of 0.05% polysorbate 80 in Ringer's lactate, and we proposed this solution for lung lavage.

On day 49, the lungs were separated with a double-lumen tube, and the patient was placed in the lateral decubitus position. After ventilation with 100% oxygen, the nondependent right lung was filled by gravity (+40 cm H²O), with the solution noted, nearly twice the lung gas volume (measured with the helium dilution technique), at a temperature of 37°C. Manual percussion of the hemithorax was performed to facilitate mixing. After 15 min, the fluid was drained. During this time, the dependent lung was ventilated with an Fio²= 0.9 and tidal volume and respiratory rate set to maintain normocapnia with a plateau pressure below 30 cm H²O. Repeated lavages were performed until the effluent solution from the lung appeared to be free of lipid. This required 15–20 procedures, and the cumulative lavage volume was approximately 20 l/lung. The lavage balance (input minus output) was close to zero. After the lavage, 250 mg pig's lung surfactant was instilled in each lobe. The entire procedure was repeated on day 50 for the left lung. Gas exchange and hemodynamics were stable during the lavage. The concentration of vaseline oil in the lung secretions on day 51 was 4 mg/ml, 10 times lower than at the start. A whole-lung CT scan on day 55 showed marked improvement in lung aeration (fig. 1). An additional lung lavage (days 67–68) did not provide any further advantage, and the oil concentration was 3.8 mg/ml.

Respiratory function improved rapidly, and the patient was successfully weaned from mechanical ventilation and discharged from the intensive care unit. After a month's rehabilitation, the patient had an oxygen fraction ratio of 371, a vital capacity 83% of predicted, a ratio of the forced expiratory volume in the first second to forced vital capacity of 82% of predicted, and a carbon dioxide diffusion capacity of 26%. A high-resolution lung CT scan showed the persistence of diffuse ground-glass opacification, more marked in the lower lobes.

Exogenous lipoid pneumonia was suspected at admission on the basis of the patient's history. She had been using one or more 200-ml packs of vaseline oil per day, even though the maximal laxative dose was 45 ml/day. Although lung injury has been described after mineral oil inhalation,5,6in this case, we could not exclude a role of intestinal absorption (normally approximately 2%7) because the homogeneous lung parenchyma alteration (fig. 1) seemed more typical of a lesion arising through the bloodstream.8The idea of lipoid pneumonia was confirmed by lipid-laden macrophages with oil drops on the surface of the fluid in the bronchoalveolar lavage and was subsequently confirmed by the coincidence of nuclear magnetic resonance and infrared spectra with those of the oil the patient had been taking.

Whatever the pathway by which the oil reaches the lung parenchyma, it either is absorbed by alveolar macrophages or remains free within the alveoli.9Because alveolar macrophages cannot metabolize it, when they die, the oil is released again into the alveoli.9In our patient, microscopic examination of fluid from sequential bronchoalveolar lavages clearly supported this because there was a cycle of intracellular oil (day 2), extracellular (day 6), intracellular (day 40), and then extracellular again. Quantitative analysis of the CT scan (day 1) showed a lung weight of 2,101.00 g with an excess tissue mass of 1,311.26 g (261.18%). The normally aerated fraction of the lung parenchyma was only 5%. In typical acute respiratory distress syndrome, these values are associated with low respiratory system compliance (< 20 ml/cm H²O),10but in this case, it was higher than expected—60 ml/cm H²O—and the lung showed an impressive opening capability (at 45 cm H²O airway pressure, the normally aerated tissue increased from 5% to 74%). This can be partially explained by the presence of vaseline oil at the gas-liquid interface acting as surfactant (surface tension: 35 dyn/cm vaseline oil, 70 dyn/cm water, 25 dyn/cm normal surfactant film).

Steroids are suggested for the treatment of lipoid pneumonia and have proved successful in some cases,5,11likely depending on the degree of intoxication. In this patient, the mineral oil concentration was 44 mg/ml. Unfortunately, no comparative data are available in the literature. At this degree of intoxication, steroids seem to control but do not solve the inflammatory response, as confirmed by the finding that after 46 days, the CT scan was similar to the initial scan, and the mineral oil was still being recycled (intracellularly and extracellularly) in the alveolar space.

We then decided on lung lavage,12preparing different solutions to remove immiscible vaseline oil. We found that possible solutions adequate to emulsify, in vitro , a mixture of saline and lung secretions in a ratio 1:1 or a mixture of saline and vaseline oil at 44 mg/ml required 40–80 mg/ml phospholipids (first solution), 625 mg/ml citicoline (second solution), or 0.5 mg/ml sorbitol monooleate (polysorbate 80) (third solution). To prepare a 40-l lavage solution (20 l/lung), the cost of the first solution would have been exorbitant, and the second solution would have resulted in hyperosmolarity. The third solution (0.5 mg/ml of polysorbate 80), however, seemed both reasonable and inexpensive. Although polysorbate 80 has apparently not been used for lung lavage in patients, it is an emulsifying agent found in several medications for enteral, parenteral, or inhalational administration13(as calyptol inhalant and fluticasone propionate), and the amount we used was below the maximal recommended daily dose (25 mg/kg).

In conclusion, this case taught us that steroids seem to control but do not solve the inflammatory response, at least for this degree of intoxication, and that lung lavage with polysorbate 80, in this patient, was safe and effective.